Results: every table¶
This page collects every table the benchmark produced, with enough context to read each one on its own. The chapters that discuss them carry the same tables beside the argument. Local measurements come from distinct saved runs on one 48 GB Mac; remote rows include network/provider latency, on one 67-query router question; chapter 5 describes the run. For the headline figures, start at Results: the short version.
Scores count correct answers out of 67. Translated means the 39 non-English queries were first translated into English; direct means the classifier got the original text. ms/query is the mean wall-clock time per query, without model loading. Chapter 5 defines every column.
Explore the interactive Pareto diagrams to find the fastest model above your own accuracy floor.
C/G labels every classifier run: C = CPU, G = GPU, C+G = mixed CPU/GPU, C+N = CPU/Neural Engine, A = automatic CPU/GPU/Neural Engine placement, R = remote. Hover over the code for execution settings and evidence.
Most tables sort when you click a column header; click again to reverse the order.
OpenRouter: nine remote models¶
All nine requested endpoints were measured on the same 67 queries, in translated and direct modes, on 3 October 2026. Translations and local results are retained from the earlier runs. Each endpoint received 106 distinct inputs, sequentially, with duplicate inputs reused across modes. Latency includes the network trip and provider processing; R means the provider did not disclose CPU/GPU placement or weight precision.
Six endpoints use a native five-way choice. Respan's three endpoints support noul only: each request contains five independent yes/no questions using the same task descriptions. We choose the largest P(true), then normalize the five values for the routing report. This noul OVR readout is labelled separately; its values are not calibrated exclusive-choice probabilities. Choose OpenRouter API in the explorer's engine filter to select all nine endpoints. Its table shows the readout for each run.
| Model | Run | Resolved version | C/G | Readout | Translated | Direct | Translated ms | Direct ms |
|---|---|---|---|---|---|---|---|---|
| typesafe/jev-1.13 | 3 October 2026 | typesafe/jev-1.13-20260917 | R | native choice | 65/67 | 65/67 | 364.4 | 375.1 |
| typesafe/jev-1.13 | earlier run | not logged | R | native choice | 65/67 | 65/67 | 466.1 | 455.2 |
| jaredpalmer/kev-4b | 3 October 2026 | jaredpalmer/kev-4b-20260924 | R | native choice | 65/67 | 65/67 | 1,653.0 | 1,525.0 |
| togethercomputer/tev1-4b-experimental | 3 October 2026 | togethercomputer/tev1-4b-experimental-20260923 | R | native choice | 64/67 | 65/67 | 339.3 | 335.1 |
| inception/mercury-decide:free | 3 October 2026 | inception/mercury-decide-20260930 | R | native choice | 64/67 | 64/67 | 408.3 | 404.5 |
| liquid/d1 | 3 October 2026 | liquid/d1-20260930 | R | native choice | 64/67 | 64/67 | 566.2 | 583.6 |
| upstage/solar-decide | 3 October 2026 | upstage/solar-decide-20260928 | R | native choice | 62/67 | 63/67 | 1,253.3 | 1,292.3 |
| respan/span-01-lite:free | 3 October 2026 | respan/span-01-lite-20260925 | R | noul OVR | 56/67 | 60/67 | 436.0 | 474.6 |
| respan/span-01-lite | 3 October 2026 | respan/span-01-lite-20260925 | R | noul OVR | 56/67 | 60/67 | 447.5 | 462.6 |
| respan/span-01 | 3 October 2026 | respan/span-01-20260925 | R | noul OVR | 56/67 | 60/67 | 487.4 | 566.3 |
All 320 methods¶
A method is one model file, read one way, on one engine or runtime. This is every method that finished, best first. Incomplete and failed attempts are left out; the latest exclusions complement the earlier failures; chapter 6 says why.
Type in the box to keep only the rows that contain every word you type: pcdServer hmm keeps the Hmm model on pcdServer, and dedicated Metal keeps the dedicated models on dohnuts (Metal).
How to read this table
- An empty direct cell means the model was trained on English only and received the translated text alone. An empty GB or load cell means the value was not measured or does not apply, for example the size of a hosted model.
- † after a method marks an execution or licence note. Hover over the method to read it.
- English counts the 28 English queries in translated mode. Other, translated and other, direct count the 39 non-English queries in each mode.
| Method | Engine | C/G | Model | Family | Quant | GB | ms/query | Translated | Direct | English | Other, translated | Other, direct | Load ms |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| openrouter@typesafe/jev-1.13 † | OpenRouter API | R | typesafe/jev-1.13 | dedicated | 364.4 | 65/67 | 65/67 | 26/28 | 39/39 | 39/39 | |||
| slot@rune-26b-a4b-q3 | slot (llama-server) | G | rune-26b-a4b | dedicated | q3 | 13.52 | 366.6 | 65/67 | 57/67 | 26/28 | 39/39 | 31/39 | 608 |
| jev † | OpenRouter API | R | typesafe/jev-1.13 | dedicated | 466.1 | 65/67 | 65/67 | 26/28 | 39/39 | 39/39 | |||
| pcdserver@rune-mradermacher-26b-a4b-q5 † | pcdServer | C+G | rune-mradermacher-26b-a4b | dedicated | q5 | 19.13 | 962.3 | 65/67 | 65/67 | 26/28 | 39/39 | 39/39 | 4,324 |
| openrouter@jaredpalmer/kev-4b † | OpenRouter API | R | jaredpalmer/kev-4b | dedicated | 1,653.0 | 65/67 | 65/67 | 26/28 | 39/39 | 39/39 | |||
| pcdserver@qwen3.5-4b-hmm-q8 | pcdServer | G | qwen3.5-4b-hmm | fine-tuned | q8 | 4.48 | 87.6 | 64/67 | 65/67 | 26/28 | 38/39 | 39/39 | 462 |
| pcdserver@qwen3.5-4b-hmm-q4 | pcdServer | G | qwen3.5-4b-hmm | fine-tuned | q4 | 2.71 | 92.6 | 64/67 | 65/67 | 26/28 | 38/39 | 39/39 | 296 |
| pcdserver@qwen3.5-4b-hmm-bf16 | pcdServer | G | qwen3.5-4b-hmm | fine-tuned | bf16 | 8.42 | 94.8 | 64/67 | 65/67 | 26/28 | 38/39 | 39/39 | 341 |
| pcdserver@rune-mradermacher-26b-a4b-q3 | pcdServer | G | rune-mradermacher-26b-a4b | dedicated | q3 | 13.29 | 146.1 | 64/67 | 64/67 | 25/28 | 39/39 | 39/39 | 354 |
| pcdserver@rune-mradermacher-26b-a4b-q5-metal † | pcdServer | G | rune-mradermacher-26b-a4b | dedicated | q5 | 19.13 | 153.3 | 64/67 | 64/67 | 25/28 | 39/39 | 39/39 | 507 |
| dohnuts-metal@decider-35b-a3b-q4 | dohnuts (Metal) | G | decider-35b-a3b | dedicated | q4 | 21.17 | 266.3 | 64/67 | 64/67 | 26/28 | 38/39 | 38/39 | |
| systemone@neohorse-4b-q8 | System One server | G | neohorse-4b | dedicated | q8 | 5.17 | 273.1 | 64/67 | 64/67 | 25/28 | 39/39 | 39/39 | |
| slot@decider-35b-a3b-q4 | slot (llama-server) | G | decider-35b-a3b | dedicated | q4 | 21.17 | 299.1 | 64/67 | 64/67 | 26/28 | 38/39 | 38/39 | 375 |
| slot@rune-26b-a4b-q4 | slot (llama-server) | G | rune-26b-a4b | dedicated | q4 | 17.03 | 335.3 | 64/67 | 55/67 | 26/28 | 38/39 | 29/39 | 439 |
| dohnuts-metal@jpt-4b-q8 † | dohnuts (Metal) | G | jpt-4b | dedicated | q8 | 4.48 | 335.6 | 64/67 | 64/67 | 25/28 | 39/39 | 39/39 | |
| openrouter@togethercomputer/tev1-4b-experimental † | OpenRouter API | R | togethercomputer/tev1-4b-experimental | dedicated | 339.3 | 64/67 | 65/67 | 26/28 | 38/39 | 39/39 | |||
| openrouter@inception/mercury-decide:free † | OpenRouter API | R | inception/mercury-decide:free | dedicated | 408.3 | 64/67 | 64/67 | 25/28 | 39/39 | 39/39 | |||
| openrouter@liquid/d1 † | OpenRouter API | R | liquid/d1 | dedicated | 566.2 | 64/67 | 64/67 | 25/28 | 39/39 | 39/39 | |||
| pcdserver@rune-mradermacher-26b-a4b-q8 † | pcdServer | C | rune-mradermacher-26b-a4b | dedicated | q8 | 26.86 | 758.1 | 64/67 | 64/67 | 25/28 | 39/39 | 39/39 | 9,856 |
| ollaya@winnow-12b | ollaya | G | winnow-12b | dedicated | q8_0 | 12.67 | 891.1 | 64/67 | 65/67 | 26/28 | 38/39 | 39/39 | 36,488 |
| gguf-head@jev-omni-q3-last-mlx † | llama.cpp + MLX head | G | jev-omni | dedicated | q3 | 6.09 | 908.4 | 64/67 | 64/67 | 26/28 | 38/39 | 38/39 | 1,193 |
| gguf-head@jev-omni-q3 | llama.cpp embeddings + head | C+G | jev-omni | dedicated | q3 | 6.09 | 1,222.3 | 64/67 | 64/67 | 26/28 | 38/39 | 38/39 | 1,411 |
| pcdserver@qwen3.5-4b-unsloth-q8 | pcdServer | G | qwen3.5-4b-unsloth | vanilla | q8 | 4.48 | 85.0 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | 346 |
| pcdserver@jpt-4b-q8 † | pcdServer | G | jpt-4b | dedicated | q8 | 4.48 | 86.3 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | 284 |
| pcdserver@jpt-4b-q4 † | pcdServer | G | jpt-4b | dedicated | q4 | 2.71 | 93.0 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | 318 |
| pcdserver@qwen3.5-4b-unsloth-q5 | pcdServer | G | qwen3.5-4b-unsloth | vanilla | q5 | 3.14 | 94.5 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | 354 |
| pcdserver@rune-mradermacher-26b-a4b-q4-metal † | pcdServer | G | rune-mradermacher-26b-a4b | dedicated | q4 | 16.80 | 124.8 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | 332 |
| pcdserver@jpt-9b-q5 † | pcdServer | G | jpt-9b | dedicated | q5 | 6.47 | 155.2 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | 495 |
| mlx@openjev-35b-a3b-4bit | MLX | G | openjev-35b-a3b | dedicated | 4bit | 20.99 | 250.0 | 63/67 | 62/67 | 24/28 | 39/39 | 38/39 | 27,426 |
| systemone@neohorse-4b-q3 | System One server | G | neohorse-4b | dedicated | q3 | 2.95 | 296.1 | 63/67 | 63/67 | 24/28 | 39/39 | 39/39 | |
| pcdserver@openjev-35b-a3b-q4 | pcdServer | G | openjev-35b-a3b | dedicated | q4 | 21.17 | 296.8 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | 338 |
| pcdserver@decider-35b-a3b-q4 | pcdServer | G | decider-35b-a3b | dedicated | q4 | 21.17 | 296.8 | 63/67 | 63/67 | 24/28 | 39/39 | 39/39 | 799 |
| slot@openjev-35b-a3b-q4 | slot (llama-server) | G | openjev-35b-a3b | dedicated | q4 | 21.17 | 311.3 | 63/67 | 62/67 | 24/28 | 39/39 | 38/39 | 388 |
| systemone@jeb-35b-a3b-q3 | System One server | G | jeb-35b-a3b | dedicated | q3 | 17.17 | 333.2 | 63/67 | 63/67 | 24/28 | 39/39 | 39/39 | |
| torch@decider-4b-bf16 | PyTorch | G | decider-4b | dedicated | bf16 | 348.2 | 63/67 | 65/67 | 26/28 | 37/39 | 39/39 | 16,802 | |
| dohnuts-metal@jpt-4b-q5 † | dohnuts (Metal) | G | jpt-4b | dedicated | q5 | 3.07 | 367.0 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | |
| pcdserver@rune-mradermacher-26b-a4b-q4 † | pcdServer | C+G | rune-mradermacher-26b-a4b | dedicated | q4 | 16.80 | 487.1 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | 3,170 |
| systemone@clef-flash-mlx4 † | Clef joint head (MLX) | G | clef-flash | dedicated | q4 | 5.28 | 554.8 | 63/67 | 63/67 | 24/28 | 39/39 | 39/39 | |
| systemone@clef-flash-q4 † | llama-server (System One) | G | clef-flash | dedicated | q4_k_m | 6.33 | 610.7 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | |
| systemone@clef-flash-q6 † | llama-server (System One) | G | clef-flash | dedicated | q6_k | 8.28 | 610.8 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | |
| dohnuts-metal@jpt-9b-q5 † | dohnuts (Metal) | G | jpt-9b | dedicated | q5 | 6.47 | 619.6 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | |
| dohnuts-metal@jpt-9b-q4 † | dohnuts (Metal) | G | jpt-9b | dedicated | q4 | 5.63 | 619.8 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | |
| dohnuts-metal@jpt-9b-q8 † | dohnuts (Metal) | G | jpt-9b | dedicated | q8 | 9.53 | 624.7 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | |
| systemone@clef-flash-q5 † | llama-server (System One) | G | clef-flash | dedicated | q5_k_m | 7.36 | 644.8 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | |
| gguf-head@jev-omni-q8 | llama.cpp embeddings + head | C+G | jev-omni | dedicated | q8 | 12.67 | 1,124.1 | 63/67 | 63/67 | 25/28 | 38/39 | 38/39 | 1,157 |
| gguf-head@jev-omni-q5 | llama.cpp embeddings + head | C+G | jev-omni | dedicated | q5 | 8.55 | 1,212.7 | 63/67 | 63/67 | 25/28 | 38/39 | 38/39 | 1,243 |
| systemone@lev-4b | System One server | G | lev-4b | dedicated | 9.49 | 2,203.8 | 63/67 | 64/67 | 25/28 | 38/39 | 39/39 | ||
| dohnuts-metal@decider-0.8b-q8 | dohnuts (Metal) | G | decider-0.8b | dedicated | q8 | 0.81 | 51.4 | 62/67 | 61/67 | 24/28 | 38/39 | 37/39 | |
| dohnuts-metal@decider-0.8b-dreamblooms-q8 | dohnuts (Metal) | G | decider-0.8b-dreamblooms | dedicated | q8 | 0.81 | 52.2 | 62/67 | 61/67 | 24/28 | 38/39 | 37/39 | |
| pcdserver@qwen2.5-3b-q8 | pcdServer | G | qwen2.5-3b | vanilla | q8 | 3.29 | 53.3 | 62/67 | 57/67 | 23/28 | 39/39 | 35/39 | 178 |
| pcdserver@qwen2.5-3b-f16 | pcdServer | G | qwen2.5-3b | vanilla | f16 | 6.18 | 53.5 | 62/67 | 60/67 | 23/28 | 39/39 | 37/39 | 252 |
| dohnuts-metal@decider-0.8b-q6 | dohnuts (Metal) | G | decider-0.8b | dedicated | q6 | 0.63 | 53.5 | 62/67 | 62/67 | 24/28 | 38/39 | 38/39 | |
| dohnuts-metal@decider-0.8b-q5 | dohnuts (Metal) | G | decider-0.8b | dedicated | q5 | 0.58 | 55.2 | 62/67 | 60/67 | 24/28 | 38/39 | 36/39 | |
| pcdserver@qwen2.5-3b-q4 | pcdServer | G | qwen2.5-3b | vanilla | q4 | 1.93 | 59.2 | 62/67 | 60/67 | 23/28 | 39/39 | 37/39 | 339 |
| pcdserver@qwen2.5-3b-q6 | pcdServer | G | qwen2.5-3b | vanilla | q6 | 2.54 | 62.0 | 62/67 | 60/67 | 23/28 | 39/39 | 37/39 | 213 |
| slot@decider-0.8b-q8 | slot (llama-server) | G | decider-0.8b | dedicated | q8 | 0.81 | 64.0 | 62/67 | 61/67 | 24/28 | 38/39 | 37/39 | 428 |
| slot@decider-0.8b-dreamblooms-q8 | slot (llama-server) | G | decider-0.8b-dreamblooms | dedicated | q8 | 0.81 | 65.1 | 62/67 | 61/67 | 24/28 | 38/39 | 37/39 | 689 |
| slot@decider-0.8b-q5 | slot (llama-server) | G | decider-0.8b | dedicated | q5 | 0.58 | 67.9 | 62/67 | 60/67 | 24/28 | 38/39 | 36/39 | 142 |
| slot@decider-0.8b-q6 | slot (llama-server) | G | decider-0.8b | dedicated | q6 | 0.63 | 68.1 | 62/67 | 62/67 | 24/28 | 38/39 | 38/39 | 216 |
| pcdserver@qwen3.5-4b-q4 | pcdServer | G | qwen3.5-4b | vanilla | q4 | 3.01 | 89.8 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | 311 |
| pcdserver@qwen3.5-4b-unsloth-q6 | pcdServer | G | qwen3.5-4b-unsloth | vanilla | q6 | 3.53 | 93.6 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | 291 |
| pcdserver@qwen3.8-4b-distill-q4 | pcdServer | G | qwen3.8-4b-distill | fine-tuned | q4 | 2.78 | 93.6 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | 284 |
| pcdserver@jpt-4b-q5 † | pcdServer | G | jpt-4b | dedicated | q5 | 3.07 | 99.4 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | 345 |
| pcdserver@qwen3.8-9b-distill-q4 | pcdServer | G | qwen3.8-9b-distill | fine-tuned | q4 | 5.78 | 137.8 | 62/67 | 64/67 | 25/28 | 37/39 | 39/39 | 449 |
| systemone@system-one-minicpm5-2b-q8 † | System One (MLX) | G | system-one-minicpm5-2b-q8 | dedicated | q8 | 2.67 | 141.8 | 62/67 | 59/67 | 24/28 | 38/39 | 35/39 | |
| dohnuts@decider-0.8b-dreamblooms-q8 | dohnuts (CPU) | C | decider-0.8b-dreamblooms | dedicated | q8 | 0.81 | 275.2 | 62/67 | 60/67 | 24/28 | 38/39 | 36/39 | |
| dohnuts-metal@jet-4b-q8 | dohnuts (Metal) | G | jet-4b | dedicated | q8 | 4.48 | 281.4 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | |
| slot@qwen3.5-4b-hmm-bf16 | slot (llama-server) | G | qwen3.5-4b-hmm | fine-tuned | bf16 | 8.42 | 284.2 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | 334 |
| slot@qwen3.5-4b-hmm-q8 | slot (llama-server) | G | qwen3.5-4b-hmm | fine-tuned | q8 | 4.48 | 287.3 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | 285 |
| dohnuts-metal@jpt-4b-q4 † | dohnuts (Metal) | G | jpt-4b | dedicated | q4 | 2.71 | 290.6 | 62/67 | 62/67 | 24/28 | 38/39 | 38/39 | |
| slot@qwen3.5-4b-hmm-q4 | slot (llama-server) | G | qwen3.5-4b-hmm | fine-tuned | q4 | 2.71 | 303.1 | 62/67 | 63/67 | 25/28 | 37/39 | 38/39 | 293 |
| splash@ornith-1.5-35b-a3b-splash † | Splash (Metal) | G | ornith-1.5-35b-a3b-splash | fine-tuned | splash-packed-q4-moe | 20.95 | 322.4 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | |
| slot@jevk5-4b-q4 | slot (llama-server) | G | jevk5-4b | dedicated | q4 | 2.71 | 350.5 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | 429 |
| systemone@imajev-4b | System One server | G | imajev-4b | dedicated | 9.81 | 363.1 | 62/67 | 64/67 | 25/28 | 37/39 | 39/39 | ||
| pcdserver@bonsai-27b-q1 | pcdServer | G | bonsai-27b | vanilla | q1 | 3.80 | 513.5 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | 1,827 |
| systemone@clef-flash-q3 † | llama-server (System One) | G | clef-flash | dedicated | q3_k_m | 4.97 | 668.3 | 62/67 | 64/67 | 25/28 | 37/39 | 39/39 | |
| kev-4b-gguf | laya.cpp | G | kev-4b | dedicated | 1,103.1 | 62/67 | 55/67 | 24/28 | 38/39 | 31/39 | |||
| openrouter@upstage/solar-decide † | OpenRouter API | R | upstage/solar-decide | dedicated | 1,253.3 | 62/67 | 63/67 | 24/28 | 38/39 | 39/39 | |||
| coreml@gliner25-decide-fp16-l256-cpu-gpu † | Core ML | C+G | gliner25-decide | dedicated | fp16 | 0.93 | 21.9 | 61/67 | 22/28 | 39/39 | 13,151 | ||
| coreai@gliner25-decide | Core AI | A | gliner25-decide | dedicated | fp16 | 0.87 | 26.9 | 61/67 | 22/28 | 39/39 | 3,378 | ||
| pcdserver@qwen3.5-2b-q3 | pcdServer | G | qwen3.5-2b | vanilla | q3 | 1.22 | 42.7 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 203 |
| dohnuts-metal@decider-0.8b-f16 | dohnuts (Metal) | G | decider-0.8b | dedicated | f16 | 1.52 | 51.9 | 61/67 | 60/67 | 23/28 | 38/39 | 37/39 | |
| dohnuts-metal@decider-0.8b-q4 | dohnuts (Metal) | G | decider-0.8b | dedicated | q4 | 0.53 | 53.3 | 61/67 | 59/67 | 23/28 | 38/39 | 36/39 | |
| pcdserver@qwen2.5-3b-q5 | pcdServer | G | qwen2.5-3b | vanilla | q5 | 2.22 | 61.7 | 61/67 | 57/67 | 22/28 | 39/39 | 36/39 | 300 |
| slot@decider-0.8b-f16 | slot (llama-server) | G | decider-0.8b | dedicated | f16 | 1.52 | 63.6 | 61/67 | 60/67 | 23/28 | 38/39 | 37/39 | 670 |
| onnx@gliner25-decide-fp32 | ONNX Runtime | C | gliner25-decide | dedicated | fp32 | 1.74 | 67.3 | 61/67 | 22/28 | 39/39 | 4,607 | ||
| slot@decider-0.8b-q4 | slot (llama-server) | G | decider-0.8b | dedicated | q4 | 0.53 | 67.4 | 61/67 | 59/67 | 23/28 | 38/39 | 36/39 | 231 |
| pcdserver@qwen3-4b-unsloth-bf16 | pcdServer | G | qwen3-4b-unsloth | vanilla | bf16 | 8.05 | 71.5 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 259 |
| pcdserver@qwen3-4b-q4 | pcdServer | G | qwen3-4b | vanilla | q4 | 2.50 | 75.5 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 230 |
| pcdserver@qwen3-4b-unsloth-q4 | pcdServer | G | qwen3-4b-unsloth | vanilla | q4 | 2.50 | 76.4 | 61/67 | 61/67 | 23/28 | 38/39 | 38/39 | 247 |
| pcdserver@qwen3.5-4b-supercoder-q4 | pcdServer | G | qwen3.5-4b-supercoder | fine-tuned | q4 | 2.61 | 84.2 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 351 |
| pcdserver@jevk5-4b-q8 | pcdServer | G | jevk5-4b | dedicated | q8 | 4.48 | 84.2 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 292 |
| executorch@decider-0.8b-fp16 | ExecuTorch (MLX) | G | decider-0.8b | dedicated | fp16 | 89.7 | 61/67 | 60/67 | 23/28 | 38/39 | 37/39 | ||
| pcdserver@jevk5-4b-q4 | pcdServer | G | jevk5-4b | dedicated | q4 | 2.71 | 89.9 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 332 |
| pcdserver@qwen3.5-4b-unsloth-iq3 | pcdServer | G | qwen3.5-4b-unsloth | vanilla | iq3 | 1.95 | 90.2 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 800 |
| pcdserver@qwen3.5-4b-unsloth-q2 | pcdServer | G | qwen3.5-4b-unsloth | vanilla | q2 | 1.94 | 93.5 | 61/67 | 61/67 | 22/28 | 39/39 | 39/39 | 829 |
| pcdserver@qwen3.5-4b-unsloth-q4 | pcdServer | G | qwen3.5-4b-unsloth | vanilla | q4 | 2.74 | 93.6 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 297 |
| pcdserver@qwen2.5-coder-7b-q4 | pcdServer | G | qwen2.5-coder-7b | fine-tuned | q4 | 4.68 | 108.9 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 386 |
| pcdserver@qwen3-8b-q4 | pcdServer | G | qwen3-8b | vanilla | q4 | 5.03 | 116.0 | 61/67 | 62/67 | 23/28 | 38/39 | 39/39 | 367 |
| pcdserver@qwen3.5-9b-unsloth-q4 | pcdServer | G | qwen3.5-9b-unsloth | vanilla | q4 | 5.68 | 136.9 | 61/67 | 63/67 | 24/28 | 37/39 | 39/39 | 452 |
| pcdserver@jpt-9b-q4 † | pcdServer | G | jpt-9b | dedicated | q4 | 5.63 | 137.3 | 61/67 | 63/67 | 24/28 | 37/39 | 39/39 | 478 |
| pcdserver@jpt-9b-q8 † | pcdServer | G | jpt-9b | dedicated | q8 | 9.53 | 141.8 | 61/67 | 63/67 | 24/28 | 37/39 | 39/39 | 451 |
| ollaya@decider-0.8b | ollaya | C | decider-0.8b | dedicated | fp32 | 1.53 | 284.8 | 61/67 | 60/67 | 23/28 | 38/39 | 37/39 | 7,924 |
| dohnuts-metal@jpt-4b-q3 † | dohnuts (Metal) | G | jpt-4b | dedicated | q3 | 2.26 | 315.0 | 61/67 | 63/67 | 24/28 | 37/39 | 39/39 | |
| slot@jevk5-4b-q8 | slot (llama-server) | G | jevk5-4b | dedicated | q8 | 4.48 | 333.9 | 61/67 | 63/67 | 24/28 | 37/39 | 39/39 | 372 |
| coreml@gliner25-decide-fp16-l256 | Core ML | A | gliner25-decide | dedicated | fp16 | 0.93 | 349.1 | 61/67 | 22/28 | 39/39 | 82,062 | ||
| dohnuts-metal@jpt-9b-q3 † | dohnuts (Metal) | G | jpt-9b | dedicated | q3 | 4.62 | 636.7 | 61/67 | 64/67 | 26/28 | 35/39 | 38/39 | |
| mlx@jev-omni-4bit | MLX | G | jev-omni | dedicated | 4bit | 6.81 | 780.1 | 61/67 | 61/67 | 24/28 | 37/39 | 37/39 | |
| pcdserver@decider-2b-f16 | pcdServer | G | decider-2b | dedicated | f16 | 3.78 | 42.2 | 60/67 | 60/67 | 22/28 | 38/39 | 38/39 | 166 |
| pcdserver@qwen2.5-coder-3b-q4 | pcdServer | G | qwen2.5-coder-3b | fine-tuned | q4 | 2.10 | 60.2 | 60/67 | 59/67 | 24/28 | 36/39 | 35/39 | 295 |
| mlx@tev1-0.8b-4bit † | MLX | G | tev1-0.8b | dedicated | 4bit | 0.42 | 64.8 | 60/67 | 59/67 | 21/28 | 39/39 | 38/39 | 3,482 |
| pcdserver@qwen3-4b-unsloth-q8 | pcdServer | G | qwen3-4b-unsloth | vanilla | q8 | 4.28 | 70.3 | 60/67 | 62/67 | 23/28 | 37/39 | 39/39 | 306 |
| pcdserver@qwen3-4b-unsloth-q3 | pcdServer | G | qwen3-4b-unsloth | vanilla | q3 | 2.08 | 78.3 | 60/67 | 61/67 | 22/28 | 38/39 | 39/39 | 706 |
| pcdserver@qwen3-4b-unsloth-q6 | pcdServer | G | qwen3-4b-unsloth | vanilla | q6 | 3.31 | 79.3 | 60/67 | 61/67 | 22/28 | 38/39 | 39/39 | 276 |
| pcdserver@qwen3-4b-unsloth-q5 | pcdServer | G | qwen3-4b-unsloth | vanilla | q5 | 2.89 | 79.4 | 60/67 | 62/67 | 23/28 | 37/39 | 39/39 | 321 |
| pcdserver@qwen3.5-4b-unsloth-q3 | pcdServer | G | qwen3.5-4b-unsloth | vanilla | q3 | 2.29 | 92.7 | 60/67 | 61/67 | 22/28 | 38/39 | 39/39 | 358 |
| pcdserver@jpt-4b-q3 † | pcdServer | G | jpt-4b | dedicated | q3 | 2.26 | 93.7 | 60/67 | 60/67 | 22/28 | 38/39 | 38/39 | 374 |
| coreml@kev-0.8b-fp16-l512 † | Core ML | C+G | kev-0.8b | dedicated | fp16 | 1.51 | 121.3 | 60/67 | 62/67 | 24/28 | 36/39 | 38/39 | 11,368 |
| pcdserver@jpt-9b-q3 † | pcdServer | G | jpt-9b | dedicated | q3 | 4.62 | 139.2 | 60/67 | 62/67 | 23/28 | 37/39 | 39/39 | 482 |
| ollaya@kev-0.8b | ollaya | C | kev-0.8b | dedicated | fp32 | 1.82 | 268.2 | 60/67 | 62/67 | 24/28 | 36/39 | 38/39 | 4,848 |
| pcdserver@decider-2b-dreamblooms-q8 | pcdServer | G | decider-2b-dreamblooms | dedicated | q8 | 2.01 | 40.1 | 59/67 | 60/67 | 22/28 | 37/39 | 38/39 | 198 |
| dohnuts-metal@tev1-0.8b-q8 † | dohnuts (Metal) | G | tev1-0.8b | dedicated | q8 | 0.81 | 61.5 | 59/67 | 62/67 | 23/28 | 36/39 | 39/39 | |
| pcdserver@qwen3-4b-unsloth-iq2 | pcdServer | G | qwen3-4b-unsloth | vanilla | iq2 | 1.53 | 72.7 | 59/67 | 57/67 | 21/28 | 38/39 | 36/39 | 605 |
| dohnuts-metal@decider-2b-f16 | dohnuts (Metal) | G | decider-2b | dedicated | f16 | 3.78 | 100.3 | 59/67 | 62/67 | 24/28 | 35/39 | 38/39 | |
| dohnuts-metal@decider-2b-dreamblooms-q8 | dohnuts (Metal) | G | decider-2b-dreamblooms | dedicated | q8 | 2.01 | 103.7 | 59/67 | 62/67 | 24/28 | 35/39 | 38/39 | |
| slot@decider-2b-f16 | slot (llama-server) | G | decider-2b | dedicated | f16 | 3.78 | 112.4 | 59/67 | 62/67 | 24/28 | 35/39 | 38/39 | 153 |
| slot@decider-2b-dreamblooms-q8 | slot (llama-server) | G | decider-2b-dreamblooms | dedicated | q8 | 2.01 | 114.4 | 59/67 | 62/67 | 24/28 | 35/39 | 38/39 | 199 |
| executorch@decider-2b-fp16 | ExecuTorch (MLX) | G | decider-2b | dedicated | fp16 | 251.3 | 59/67 | 62/67 | 24/28 | 35/39 | 38/39 | ||
| pcdserver@decider-0.8b-q5 | pcdServer | G | decider-0.8b | dedicated | q5 | 0.58 | 29.9 | 57/67 | 57/67 | 22/28 | 36/39 | 35/39 | 180 |
| pcdserver@qwen3.5-2b-q8 | pcdServer | G | qwen3.5-2b | vanilla | q8 | 2.08 | 39.5 | 57/67 | 59/67 | 20/28 | 38/39 | 39/39 | 189 |
| pcdserver@qwen3.5-2b-unsloth-q8 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | q8 | 2.01 | 39.5 | 57/67 | 59/67 | 20/28 | 38/39 | 39/39 | 126 |
| pcdserver@decider-0.8b-q4 | pcdServer | G | decider-0.8b | dedicated | q4 | 0.53 | 40.2 | 57/67 | 57/67 | 22/28 | 36/39 | 35/39 | 583 |
| pcdserver@decider-2b-iq4 | pcdServer | G | decider-2b | dedicated | iq4 | 1.24 | 41.4 | 57/67 | 57/67 | 21/28 | 37/39 | 37/39 | 167 |
| pcdserver@qwen3.5-2b-unsloth-bf16 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | bf16 | 3.78 | 42.2 | 57/67 | 59/67 | 20/28 | 38/39 | 39/39 | 328 |
| pcdserver@qwen3.5-2b-unsloth-q4 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | q4 | 1.28 | 44.8 | 57/67 | 57/67 | 19/28 | 39/39 | 38/39 | 139 |
| dohnuts-metal@decider-0.8b-q3 | dohnuts (Metal) | G | decider-0.8b | dedicated | q3 | 0.47 | 54.5 | 57/67 | 55/67 | 24/28 | 34/39 | 31/39 | |
| pcdserver@qwen2.5-3b-q3 | pcdServer | G | qwen2.5-3b | vanilla | q3 | 1.59 | 60.6 | 57/67 | 51/67 | 20/28 | 38/39 | 31/39 | 672 |
| ollaya@von-1.1 | ollaya | C | von-1.1 | dedicated | fp32 | 1.59 | 62.8 | 57/67 | 22/28 | 36/39 | 14,254 | ||
| slot@decider-0.8b-q3 | slot (llama-server) | G | decider-0.8b | dedicated | q3 | 0.47 | 69.7 | 57/67 | 55/67 | 24/28 | 34/39 | 31/39 | 101 |
| dohnuts-metal@this-that-model-1.2-q3 | dohnuts (Metal) | G | this-that-model-1.2 | dedicated | q3 | 1.10 | 113.1 | 57/67 | 56/67 | 23/28 | 35/39 | 33/39 | |
| systemone@leo-1.7b | System One server | G | leo-1.7b | dedicated | 3.52 | 136.2 | 57/67 | 62/67 | 23/28 | 35/39 | 39/39 | ||
| systemone@kev-ggmlc-0.8b-q8 † | ggmlc (Metal) | G | kev-0.8b | dedicated | q8 | 0.83 | 199.8 | 57/67 | 45/67 | 24/28 | 34/39 | 21/39 | |
| kev-0.8b-gguf | laya.cpp | G | kev-0.8b | dedicated | 231.5 | 57/67 | 45/67 | 24/28 | 34/39 | 21/39 | |||
| pcdserver@qwen3.5-0.8b-q5 | pcdServer | G | qwen3.5-0.8b | vanilla | q5 | 0.65 | 32.1 | 57/67 | 57/67 | 21/28 | 36/39 | 37/39 | 284 |
| pcdserver@qwen3-1.7b-unsloth-q8 | pcdServer | G | qwen3-1.7b-unsloth | vanilla | q8 | 1.83 | 35.1 | 57/67 | 59/67 | 21/28 | 36/39 | 38/39 | 95 |
| pcdserver@qwen3-1.7b-q8 | pcdServer | G | qwen3-1.7b | vanilla | q8 | 2.17 | 35.2 | 57/67 | 59/67 | 21/28 | 36/39 | 38/39 | 94 |
| pcdserver@qwen3-1.7b-bf16 | pcdServer | G | qwen3-1.7b | vanilla | bf16 | 4.07 | 36.2 | 57/67 | 59/67 | 21/28 | 36/39 | 38/39 | 206 |
| pcdserver@qwen3-1.7b-q5 | pcdServer | G | qwen3-1.7b | vanilla | q5 | 1.47 | 38.7 | 57/67 | 59/67 | 21/28 | 36/39 | 38/39 | 148 |
| pcdserver@qwen3-1.7b-unsloth-q5 | pcdServer | G | qwen3-1.7b-unsloth | vanilla | q5 | 1.26 | 38.7 | 57/67 | 59/67 | 21/28 | 36/39 | 38/39 | 170 |
| pcdserver@qwen3-1.7b-unsloth-q6 | pcdServer | G | qwen3-1.7b-unsloth | vanilla | q6 | 1.42 | 39.2 | 57/67 | 59/67 | 21/28 | 36/39 | 38/39 | 158 |
| pcdserver@qwen3-1.7b-q6 | pcdServer | G | qwen3-1.7b | vanilla | q6 | 1.67 | 39.3 | 57/67 | 59/67 | 21/28 | 36/39 | 38/39 | 150 |
| pcdserver@qwen3.5-2b-unsloth-iq3 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | iq3 | 0.93 | 40.2 | 57/67 | 57/67 | 21/28 | 36/39 | 37/39 | 435 |
| pcdserver@qwen3.5-2b-unsloth-q3 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | q3 | 1.11 | 41.9 | 57/67 | 59/67 | 20/28 | 37/39 | 39/39 | 165 |
| pcdserver@qwen3.5-2b-q5 | pcdServer | G | qwen3.5-2b | vanilla | q5 | 1.57 | 42.3 | 57/67 | 57/67 | 19/28 | 38/39 | 39/39 | 196 |
| pcdserver@qwen3.5-2b-q6 | pcdServer | G | qwen3.5-2b | vanilla | q6 | 1.70 | 42.6 | 57/67 | 57/67 | 19/28 | 38/39 | 39/39 | 134 |
| pcdserver@qwen3.5-2b-unsloth-q5 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | q5 | 1.44 | 43.5 | 57/67 | 57/67 | 19/28 | 38/39 | 38/39 | 142 |
| pcdserver@qwen3.5-2b-q4 | pcdServer | G | qwen3.5-2b | vanilla | q4 | 1.40 | 47.2 | 57/67 | 57/67 | 19/28 | 38/39 | 39/39 | 217 |
| onnx@von-1.2 | ONNX Runtime | C | von-1.2 | dedicated | fp32 | 1.58 | 52.6 | 57/67 | 22/28 | 35/39 | 2,485 | ||
| ollaya@decision-eos | ollaya | C | decision-eos | dedicated | fp32 | 1.54 | 322.9 | 57/67 | 56/67 | 23/28 | 34/39 | 33/39 | 5,638 |
| pcdserver@qwen3.5-0.8b-unsloth-q5 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | q5 | 0.59 | 29.6 | 56/67 | 57/67 | 21/28 | 35/39 | 36/39 | 184 |
| pcdserver@qwen3-1.7b-unsloth-q4 | pcdServer | G | qwen3-1.7b-unsloth | vanilla | q4 | 1.11 | 40.2 | 56/67 | 57/67 | 20/28 | 36/39 | 38/39 | 194 |
| pcdserver@qwen3.5-2b-unsloth-q6 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | q6 | 1.57 | 43.8 | 56/67 | 57/67 | 19/28 | 37/39 | 39/39 | 132 |
| dohnuts-metal@decider-2b-iq4 | dohnuts (Metal) | G | decider-2b | dedicated | iq4 | 1.24 | 106.7 | 56/67 | 57/67 | 23/28 | 33/39 | 35/39 | |
| slot@decider-2b-iq4 | slot (llama-server) | G | decider-2b | dedicated | iq4 | 1.24 | 126.8 | 56/67 | 57/67 | 23/28 | 33/39 | 35/39 | 719 |
| openrouter@respan/span-01-lite:free † | OpenRouter API | R | respan/span-01-lite:free | dedicated | 436.0 | 56/67 | 60/67 | 21/28 | 35/39 | 39/39 | |||
| openrouter@respan/span-01-lite † | OpenRouter API | R | respan/span-01-lite | dedicated | 447.5 | 56/67 | 60/67 | 21/28 | 35/39 | 39/39 | |||
| openrouter@respan/span-01 † | OpenRouter API | R | respan/span-01 | dedicated | 487.4 | 56/67 | 60/67 | 21/28 | 35/39 | 39/39 | |||
| pcdserver@decider-0.8b-dreamblooms-q8 | pcdServer | G | decider-0.8b-dreamblooms | dedicated | q8 | 0.81 | 28.3 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 78 |
| pcdserver@qwen3.5-0.8b-unsloth-q8 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | q8 | 0.81 | 28.7 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 177 |
| pcdserver@qwen3.5-0.8b-q6 | pcdServer | G | qwen3.5-0.8b | vanilla | q6 | 0.69 | 29.0 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 450 |
| pcdserver@qwen3.5-0.8b-unsloth-q6 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | q6 | 0.64 | 29.2 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 186 |
| pcdserver@decider-0.8b-q6 | pcdServer | G | decider-0.8b | dedicated | q6 | 0.63 | 29.8 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 280 |
| pcdserver@qwen3.5-0.8b-bf16 | pcdServer | G | qwen3.5-0.8b | vanilla | bf16 | 1.56 | 30.0 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 407 |
| pcdserver@decider-0.8b-q8 | pcdServer | G | decider-0.8b | dedicated | q8 | 0.81 | 30.2 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 208 |
| pcdserver@qwen3.5-0.8b-ggml-bf16 | pcdServer | G | qwen3.5-0.8b-ggml | vanilla | bf16 | 1.56 | 30.3 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 337 |
| pcdserver@decider-0.8b-f16 | pcdServer | G | decider-0.8b | dedicated | f16 | 1.52 | 30.7 | 55/67 | 56/67 | 19/28 | 36/39 | 37/39 | 438 |
| pcdserver@qwen3.5-0.8b-unsloth-bf16 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | bf16 | 1.52 | 31.0 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 208 |
| pcdserver@qwen3.5-0.8b-q8 | pcdServer | G | qwen3.5-0.8b | vanilla | q8 | 0.83 | 31.2 | 55/67 | 57/67 | 20/28 | 35/39 | 37/39 | 252 |
| pcdserver@qwen3.5-0.8b-unsloth-q4 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | q4 | 0.53 | 50.1 | 55/67 | 54/67 | 20/28 | 35/39 | 34/39 | 255 |
| systemone@bosun-1.7b-q5 † | llama-cpp-python (Metal) | G | bosun-1.7b | dedicated | q5 | 1.26 | 130.8 | 55/67 | 55/67 | 22/28 | 33/39 | 33/39 | |
| pcdserver@pulse-0.6b-q4 | pcdServer | G | pulse-0.6b | fine-tuned | q4 | 0.40 | 27.9 | 54/67 | 52/67 | 21/28 | 33/39 | 31/39 | 196 |
| pcdserver@bonsai-1.7b-q2 | pcdServer | G | bonsai-1.7b | vanilla | q2 | 0.49 | 34.9 | 54/67 | 55/67 | 19/28 | 35/39 | 36/39 | 259 |
| pcdserver@qwen3-1.7b-q2 | pcdServer | G | qwen3-1.7b | vanilla | q2 | 0.88 | 37.5 | 54/67 | 50/67 | 21/28 | 33/39 | 29/39 | 104 |
| pcdserver@qwen3-1.7b-q4 | pcdServer | G | qwen3-1.7b | vanilla | q4 | 1.28 | 40.7 | 54/67 | 57/67 | 20/28 | 34/39 | 38/39 | 343 |
| gguf-head@laya-neutron-ml | llama.cpp embeddings + head | C+G | laya-multilingual | dedicated | q4_0 | 0.27 | 42.7 | 54/67 | 51/67 | 21/28 | 33/39 | 30/39 | 2,739 |
| gguf-head@laya-neutron-ml-mps † | llama.cpp + MPS head | G | laya-multilingual | dedicated | q4_0 | 0.27 | 50.3 | 54/67 | 51/67 | 21/28 | 33/39 | 30/39 | 4,915 |
| dohnuts-metal@this-that-model-1.2-q5 | dohnuts (Metal) | G | this-that-model-1.2 | dedicated | q5 | 1.41 | 117.4 | 54/67 | 57/67 | 21/28 | 33/39 | 37/39 | |
| systemone@kev-ggmlc-0.8b-q4 † | ggmlc (Metal) | G | kev-0.8b | dedicated | q4 | 0.96 | 207.2 | 54/67 | 43/67 | 23/28 | 31/39 | 20/39 | |
| pcdserver@qwen3-0.6b-q8 | pcdServer | G | qwen3-0.6b | vanilla | q8 | 0.64 | 25.5 | 53/67 | 54/67 | 18/28 | 35/39 | 36/39 | 191 |
| pcdserver@qwen3.5-0.8b-ggml-q4 | pcdServer | G | qwen3.5-0.8b-ggml | vanilla | q4 | 0.56 | 28.8 | 53/67 | 55/67 | 19/28 | 34/39 | 36/39 | 801 |
| pcdserver@qwen3-1.7b-unsloth-q3 | pcdServer | G | qwen3-1.7b-unsloth | vanilla | q3 | 0.94 | 37.6 | 53/67 | 56/67 | 18/28 | 35/39 | 38/39 | 216 |
| pcdserver@qwen3.5-2b-unsloth-q2 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | q2 | 0.97 | 41.3 | 53/67 | 48/67 | 19/28 | 34/39 | 29/39 | 255 |
| mlx@laya-ml-mxfp8 † | MLX | G | laya-multilingual | dedicated | mxfp8 | 0.32 | 57.8 | 53/67 | 49/67 | 20/28 | 33/39 | 29/39 | 2,639 |
| pcdserver@qwen2.5-3b-q2 | pcdServer | G | qwen2.5-3b | vanilla | q2 | 1.27 | 58.3 | 53/67 | 46/67 | 17/28 | 36/39 | 29/39 | 271 |
| dohnuts-metal@this-that-model-1.2-q8 | dohnuts (Metal) | G | this-that-model-1.2 | dedicated | q8 | 2.01 | 107.8 | 53/67 | 57/67 | 20/28 | 33/39 | 38/39 | |
| systemone@system-one-scorer-onnx † | ONNX scalar (CPU) | C | system-one-scorer-onnx | dedicated | q4 | 2.64 | 1,410.8 | 53/67 | 23/28 | 30/39 | |||
| laya-mlx | MLX | G | laya-multilingual | dedicated | 6.9 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 1,797 | ||
| laya-coreai | Core AI | A | laya-multilingual | dedicated | 10.0 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 3,164 | ||
| systemone@laya-mlx-package-fp16 † | Laya MLX (native) | G | laya-multilingual | dedicated | fp16 | 0.64 | 11.4 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | |
| ollaya@laya-multilingual | ollaya | G | laya-multilingual | dedicated | fp32 | 0.68 | 12.3 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 1,217 |
| laya-coreai-mlb | Core AI | A | laya-multilingual | dedicated | 19.4 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 2,404 | ||
| laya-ml-onnx-fp32 | ONNX Runtime | C | laya-multilingual | dedicated | fp32 | 23.4 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 825 | |
| onnx@laya-ml-fp32 | ONNX Runtime | C | laya-multilingual | dedicated | fp32 | 1.29 | 24.7 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 1,578 |
| laya-ml-onnx-int8 | ONNX Runtime | C | laya-multilingual | dedicated | int8 | 25.5 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 741 | |
| pcdserver@decider-0.8b-q3 | pcdServer | G | decider-0.8b | dedicated | q3 | 0.47 | 29.3 | 52/67 | 47/67 | 18/28 | 34/39 | 29/39 | 161 |
| pcdserver@qwen3.5-0.8b-q4 | pcdServer | G | qwen3.5-0.8b | vanilla | q4 | 0.58 | 30.2 | 52/67 | 51/67 | 17/28 | 35/39 | 34/39 | 258 |
| laya-wd-gguf-q8 | llama.cpp embeddings | C+G | laya-multilingual | dedicated | q8 | 44.4 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 2,337 | |
| laya-wd-gguf-f16 | llama.cpp embeddings | C+G | laya-multilingual | dedicated | f16 | 44.4 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 2,437 | |
| laya-wd-gguf-q8-mps † | llama.cpp + MPS head | G | laya-multilingual | dedicated | q8 | 0.37 | 45.9 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 86,773 |
| laya-ml-gguf | laya.cpp | C | laya-multilingual | dedicated | 57.6 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | |||
| onnx@laya-ml-fp16 | ONNX Runtime | C | laya-multilingual | dedicated | fp16 | 0.65 | 57.6 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | 1,067 |
| laya-ml-gguf-f16 | laya.cpp | C | laya-multilingual | dedicated | f16 | 61.4 | 52/67 | 49/67 | 21/28 | 31/39 | 28/39 | ||
| pcdserver@qwen3.5-4b-unsloth-iq2 | pcdServer | G | qwen3.5-4b-unsloth | vanilla | iq2 | 1.52 | 87.0 | 52/67 | 54/67 | 16/28 | 36/39 | 38/39 | 659 |
| laya-wd-gguf-q6 | llama.cpp embeddings | C+G | laya-multilingual | dedicated | q6 | 44.5 | 51/67 | 47/67 | 20/28 | 31/39 | 27/39 | 2,276 | |
| laya-ml-gguf-q4 | laya.cpp | C | laya-multilingual | dedicated | q4 | 55.0 | 51/67 | 47/67 | 18/28 | 33/39 | 29/39 | ||
| dohnuts-metal@this-that-model-1.2-q4 | dohnuts (Metal) | G | this-that-model-1.2 | dedicated | q4 | 1.27 | 110.5 | 51/67 | 51/67 | 19/28 | 32/39 | 32/39 | |
| mlx@semif-4b-nli | MLX | G | semif-4b-nli | dedicated | 8.41 | 120.6 | 51/67 | 49/67 | 19/28 | 32/39 | 30/39 | 13,083 | |
| pcdserver@pulse-0.6b-q8 | pcdServer | G | pulse-0.6b | fine-tuned | q8 | 0.64 | 23.2 | 50/67 | 47/67 | 20/28 | 30/39 | 27/39 | 56 |
| pcdserver@lq-decide-1.7b-q8 | pcdServer | G | lq-decide-1.7b | fine-tuned | q8 | 1.83 | 35.1 | 50/67 | 46/67 | 20/28 | 30/39 | 26/39 | 116 |
| pcdserver@qwen3-1.7b-q3 | pcdServer | G | qwen3-1.7b | vanilla | q3 | 1.07 | 37.7 | 50/67 | 53/67 | 16/28 | 34/39 | 37/39 | 109 |
| pcdserver@qwen3.5-2b-q2 | pcdServer | G | qwen3.5-2b | vanilla | q2 | 1.07 | 41.5 | 49/67 | 49/67 | 19/28 | 30/39 | 30/39 | 127 |
| pcdserver@qwen3.8-2b-distill-q4 | pcdServer | G | qwen3.8-2b-distill | fine-tuned | q4 | 1.31 | 42.9 | 49/67 | 51/67 | 19/28 | 30/39 | 32/39 | 154 |
| laya-wd-gguf-q3 | llama.cpp embeddings | C+G | laya-multilingual | dedicated | q3 | 45.3 | 49/67 | 43/67 | 17/28 | 32/39 | 26/39 | 3,284 | |
| systemone@system-one-qwen3-0.6b † | System One (MLX) | G | system-one-qwen3-0.6b | dedicated | bf16+adapter | 1.52 | 47.7 | 49/67 | 52/67 | 21/28 | 28/39 | 31/39 | |
| pcdserver@qwen2.5-3b-iq2 | pcdServer | G | qwen2.5-3b | vanilla | iq2 | 1.14 | 55.5 | 49/67 | 49/67 | 16/28 | 33/39 | 33/39 | 427 |
| laya-coreml-ane-w8-compact | Core ML, Neural Engine | C+N | laya-multilingual | dedicated | w8 | 4.5 | 48/67 | 49/67 | 17/28 | 31/39 | 32/39 | 21,052 | |
| laya-typed-decisions | PyTorch | G | laya-typed-decisions | dedicated | 17.7 | 48/67 | 43/67 | 18/28 | 30/39 | 25/39 | 4,793 | ||
| pcdserver@qwen3.5-0.8b-unsloth-q3 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | q3 | 0.47 | 30.1 | 48/67 | 48/67 | 17/28 | 31/39 | 31/39 | 186 |
| laya-coreml-ane-compact | Core ML, Neural Engine | C+N | laya-multilingual | dedicated | 4.2 | 47/67 | 49/67 | 17/28 | 30/39 | 32/39 | 20,130 | ||
| laya-mlx-compact | MLX | G | laya-multilingual | dedicated | 7.9 | 47/67 | 49/67 | 17/28 | 30/39 | 32/39 | 11,010 | ||
| pcdserver@bonsai-1.7b-q1 | pcdServer | G | bonsai-1.7b | vanilla | q1 | 0.25 | 34.9 | 47/67 | 41/67 | 14/28 | 33/39 | 27/39 | 243 |
| laya-wd-gguf-q5 | llama.cpp embeddings | C+G | laya-multilingual | dedicated | q5 | 45.5 | 47/67 | 43/67 | 17/28 | 30/39 | 26/39 | 2,350 | |
| pcdserver@qwen3.5-0.8b-q3 | pcdServer | G | qwen3.5-0.8b | vanilla | q3 | 0.52 | 31.3 | 46/67 | 40/67 | 15/28 | 31/39 | 25/39 | 218 |
| gguf-head@laya-neutron-en | llama.cpp embeddings + head | C+G | laya-en | dedicated | q4_0 | 0.29 | 62.4 | 46/67 | 16/28 | 30/39 | 1,664 | ||
| systemone@xdecision-q8-mlx † | xDecision (MLX) | G | xDecision | dedicated | q8 | 0.40 | 14.3 | 45/67 | 43/67 | 17/28 | 28/39 | 26/39 | |
| pcdserver@qwen3.8-2b-distill-q5 | pcdServer | G | qwen3.8-2b-distill | fine-tuned | q5 | 1.45 | 46.1 | 45/67 | 45/67 | 17/28 | 28/39 | 28/39 | 140 |
| systemone@xdecision-f16-mlx † | xDecision (MLX) | G | xDecision | dedicated | f16 | 0.70 | 14.9 | 44/67 | 43/67 | 16/28 | 28/39 | 27/39 | |
| pcdserver@qwen3-1.7b-unsloth-q2 | pcdServer | G | qwen3-1.7b-unsloth | vanilla | q2 | 0.78 | 37.4 | 44/67 | 46/67 | 18/28 | 26/39 | 28/39 | 276 |
| pcdserver@qwen3.8-2b-distill-q8 | pcdServer | G | qwen3.8-2b-distill | fine-tuned | q8 | 2.08 | 39.5 | 44/67 | 45/67 | 17/28 | 27/39 | 28/39 | 129 |
| pcdserver@qwen3.5-2b-unsloth-iq2 | pcdServer | G | qwen3.5-2b-unsloth | vanilla | iq2 | 0.77 | 40.8 | 44/67 | 39/67 | 15/28 | 29/39 | 24/39 | 129 |
| mlx@laya-ml-mxfp4 † | MLX | G | laya-multilingual | dedicated | mxfp4 | 0.16 | 57.9 | 44/67 | 44/67 | 16/28 | 28/39 | 28/39 | 2,438 |
| pcdserver@qwen3.5-0.8b-unsloth-q2 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | q2 | 0.42 | 28.5 | 43/67 | 42/67 | 14/28 | 29/39 | 28/39 | 764 |
| pcdserver@qwen3.8-2b-distill-q6 | pcdServer | G | qwen3.8-2b-distill | fine-tuned | q6 | 1.61 | 43.5 | 43/67 | 43/67 | 16/28 | 27/39 | 27/39 | 122 |
| ollaya@laya-en | ollaya | G | laya-en | dedicated | fp32 | 0.85 | 19.4 | 42/67 | 16/28 | 26/39 | 1,804 | ||
| systemone@system-one-distilled † | Transformers NLI (MPS) | G | system-one-distilled | dedicated | fp32 | 0.28 | 21.2 | 42/67 | 17/28 | 25/39 | |||
| systemone@system-one-gold † | Transformers NLI (MPS) | G | system-one-gold | dedicated | fp32 | 0.28 | 21.4 | 42/67 | 17/28 | 25/39 | |||
| pcdserver@zerocoder-0.8b-q2 | pcdServer | G | zerocoder-0.8b | fine-tuned | q2 | 0.38 | 34.7 | 42/67 | 36/67 | 15/28 | 27/39 | 21/39 | 339 |
| systemone@raz-nli-xsmall-openjev † | Raz NLI (MPS) | G | raz-nli-xsmall-openjev | dedicated | fp32 | 0.28 | 49.4 | 42/67 | 14/28 | 28/39 | |||
| pcdserver@qwen1.5-4b-q8 | pcdServer | G | qwen1.5-4b | vanilla | q8 | 4.20 | 247.8 | 42/67 | 43/67 | 16/28 | 26/39 | 27/39 | 322 |
| pcdserver@lq-decide-1.7b-q4 | pcdServer | G | lq-decide-1.7b | fine-tuned | q4 | 1.11 | 43.8 | 41/67 | 42/67 | 15/28 | 26/39 | 27/39 | 108 |
| pcdserver@qwen1.5-4b-q4 | pcdServer | G | qwen1.5-4b | vanilla | q4 | 2.46 | 232.1 | 41/67 | 43/67 | 16/28 | 25/39 | 27/39 | 391 |
| pcdserver@qwen1.5-4b-q2 | pcdServer | G | qwen1.5-4b | vanilla | q2 | 1.62 | 233.0 | 41/67 | 42/67 | 16/28 | 25/39 | 26/39 | 444 |
| pcdserver@qwen3-reranker-0.6b-q5 | pcdServer | G | qwen3-reranker-0.6b | fine-tuned | q5 | 0.44 | 24.7 | 40/67 | 44/67 | 15/28 | 25/39 | 29/39 | 58 |
| pcdserver@qwen3.5-0.8b-q2 | pcdServer | G | qwen3.5-0.8b | vanilla | q2 | 0.46 | 32.2 | 40/67 | 42/67 | 13/28 | 27/39 | 29/39 | 541 |
| pcdserver@qwen1.5-4b-q6 | pcdServer | G | qwen1.5-4b | vanilla | q6 | 3.25 | 275.7 | 40/67 | 43/67 | 16/28 | 24/39 | 27/39 | 309 |
| systemone@clm-8b-q8 | System One server | C+G | clm-8b | dedicated | q8 | 8.25 | 107.6 | 39/67 | 28/67 | 14/28 | 25/39 | 15/39 | |
| pcdserver@qwen1.5-4b-q3 | pcdServer | G | qwen1.5-4b | vanilla | q3 | 2.03 | 237.2 | 39/67 | 43/67 | 16/28 | 23/39 | 27/39 | 441 |
| pcdserver@qwen1.5-4b-q5 | pcdServer | G | qwen1.5-4b | vanilla | q5 | 2.84 | 283.6 | 39/67 | 42/67 | 16/28 | 23/39 | 26/39 | 424 |
| systemone@clm-8b-mlx6 † | CLM embeddings + head (MLX/CPU) | C+G | clm-8b | dedicated | q6 | 7.24 | 415.7 | 39/67 | 14/28 | 25/39 | |||
| ollaya@clm-8b | ollaya | C | clm-8b | dedicated | bf16 | 16.50 | 1,677.9 | 39/67 | 30/67 | 14/28 | 25/39 | 16/39 | 49,184 |
| systemone@clm-8b-mlx8 | System One server | C+G | clm-8b | dedicated | mlx8 | 9.29 | 100.4 | 38/67 | 31/67 | 14/28 | 24/39 | 17/39 | |
| systemone@clm-8b-q5 | System One server | C+G | clm-8b | dedicated | q5 | 5.78 | 119.2 | 38/67 | 27/67 | 13/28 | 25/39 | 14/39 | |
| systemone@clm-8b-q6-k † | CLM embeddings + head (Metal/CPU) | C+G | clm-8b | dedicated | q6_k | 7.10 | 490.7 | 38/67 | 13/28 | 25/39 | |||
| systemone@clm-8b-q5-km † | CLM embeddings + head (Metal/CPU) | C+G | clm-8b | dedicated | q5_k_m | 6.31 | 509.9 | 38/67 | 13/28 | 25/39 | |||
| onnx@decima-small-int8 | ONNX Runtime | C | decima-small | dedicated | int8 | 0.13 | 12.4 | 37/67 | 38/67 | 13/28 | 24/39 | 25/39 | 2,290 |
| systemone@system-one-zeroshot † | Transformers NLI (MPS) | G | system-one-zeroshot | dedicated | fp32 | 0.28 | 21.2 | 37/67 | 14/28 | 23/39 | |||
| laya-wd-gguf-q4 | llama.cpp embeddings | C+G | laya-multilingual | dedicated | q4 | 43.3 | 36/67 | 38/67 | 14/28 | 22/39 | 24/39 | 2,243 | |
| systemone@clm-8b-q4 | System One server | C+G | clm-8b | dedicated | q4 | 5.03 | 113.8 | 36/67 | 28/67 | 13/28 | 23/39 | 16/39 | |
| systemone@clm-8b-q4-km † | CLM embeddings + head (Metal/CPU) | C+G | clm-8b | dedicated | q4_k_m | 5.57 | 449.4 | 36/67 | 13/28 | 23/39 | |||
| pcdserver@qwen1.5-0.5b-q3 | pcdServer | G | qwen1.5-0.5b | vanilla | q3 | 0.35 | 21.2 | 35/67 | 32/67 | 11/28 | 24/39 | 21/39 | 248 |
| pcdserver@lyco-router-0.6b-q4 | pcdServer | G | lyco-router-0.6b | fine-tuned | q4 | 0.48 | 26.4 | 35/67 | 33/67 | 12/28 | 23/39 | 21/39 | 514 |
| torch@pulse-decide-150m † | PyTorch | G | pulse-decide-150m | dedicated | 0.60 | 17.9 | 34/67 | 15/28 | 19/39 | 10,043 | |||
| pcdserver@qwen3.5-0.8b-unsloth-iq3 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | iq3 | 0.40 | 28.2 | 34/67 | 31/67 | 12/28 | 22/39 | 19/39 | 246 |
| pcdserver@qwen1.5-1.8b-q3 | pcdServer | G | qwen1.5-1.8b | vanilla | q3 | 1.02 | 41.8 | 34/67 | 35/67 | 15/28 | 19/39 | 20/39 | 510 |
| pcdserver@qwen3-reranker-0.6b-q6 | pcdServer | G | qwen3-reranker-0.6b | fine-tuned | q6 | 0.49 | 24.5 | 33/67 | 37/67 | 13/28 | 20/39 | 24/39 | 61 |
| pcdserver@qwen1.5-1.8b-q4 | pcdServer | G | qwen1.5-1.8b | vanilla | q4 | 1.22 | 40.0 | 33/67 | 39/67 | 15/28 | 18/39 | 24/39 | 343 |
| torch@lumma-fev-0.6b | PyTorch | G | lumma-fev-0.6b | dedicated | 1.30 | 69.5 | 33/67 | 28/67 | 14/28 | 19/39 | 14/39 | 6,272 | |
| systemone@clm-8b-mlx4 | System One server | C+G | clm-8b | dedicated | mlx4 | 5.20 | 89.6 | 33/67 | 28/67 | 11/28 | 22/39 | 18/39 | |
| pcdserver@zerocoder-0.8b-q3 | pcdServer | G | zerocoder-0.8b | fine-tuned | q3 | 0.45 | 34.6 | 32/67 | 30/67 | 11/28 | 21/39 | 19/39 | 291 |
| pcdserver@qwen1.5-1.8b-q8 | pcdServer | G | qwen1.5-1.8b | vanilla | q8 | 1.96 | 36.4 | 32/67 | 36/67 | 13/28 | 19/39 | 23/39 | 106 |
| pcdserver@qwen1.5-1.8b-q5 | pcdServer | G | qwen1.5-1.8b | vanilla | q5 | 1.38 | 41.3 | 32/67 | 35/67 | 13/28 | 19/39 | 22/39 | 327 |
| mlx@julia-1 | MLX | G | julia-1 | dedicated | 0.58 | 6.8 | 31/67 | 31/67 | 11/28 | 20/39 | 20/39 | 1,866 | |
| onnx@julia-1 | ONNX Runtime | C | julia-1 | dedicated | fp32 | 0.58 | 12.2 | 31/67 | 31/67 | 11/28 | 20/39 | 20/39 | 976 |
| pcdserver@qwen1.5-1.8b-q6 | pcdServer | G | qwen1.5-1.8b | vanilla | q6 | 1.58 | 39.8 | 31/67 | 37/67 | 14/28 | 17/39 | 23/39 | 334 |
| pcdserver@qwen1.5-0.5b-q6 | pcdServer | G | qwen1.5-0.5b | vanilla | q6 | 0.51 | 22.6 | 30/67 | 28/67 | 8/28 | 22/39 | 21/39 | 197 |
| pcdserver@qwen3-reranker-0.6b-q8 | pcdServer | G | qwen3-reranker-0.6b | fine-tuned | q8 | 0.64 | 23.3 | 30/67 | 33/67 | 12/28 | 18/39 | 21/39 | 57 |
| pcdserver@qwen3-reranker-0.6b-f16 | pcdServer | G | qwen3-reranker-0.6b | fine-tuned | f16 | 1.20 | 23.8 | 30/67 | 32/67 | 12/28 | 18/39 | 20/39 | 53 |
| pcdserver@qwen1.5-0.5b-q8 | pcdServer | G | qwen1.5-0.5b | vanilla | q8 | 0.66 | 21.1 | 28/67 | 28/67 | 7/28 | 22/39 | 21/39 | 177 |
| pcdserver@qwen1.5-0.5b-q5 | pcdServer | G | qwen1.5-0.5b | vanilla | q5 | 0.46 | 21.7 | 28/67 | 28/67 | 7/28 | 22/39 | 22/39 | 238 |
| pcdserver@qwen2.5-0.5b-q4 | pcdServer | G | qwen2.5-0.5b | vanilla | q4 | 0.49 | 20.2 | 28/67 | 26/67 | 10/28 | 18/39 | 16/39 | 605 |
| pcdserver@qwen3-reranker-0.6b-q4 | pcdServer | G | qwen3-reranker-0.6b | fine-tuned | q4 | 0.40 | 29.2 | 28/67 | 34/67 | 10/28 | 18/39 | 24/39 | 339 |
| pcdserver@qwen2.5-0.5b-pcb-q2 | pcdServer | G | qwen2.5-0.5b-pcb | fine-tuned | q2 | 0.34 | 18.1 | 27/67 | 23/67 | 9/28 | 18/39 | 14/39 | 334 |
| pcdserver@zerocoder-0.8b-f16 | pcdServer | G | zerocoder-0.8b | fine-tuned | f16 | 1.64 | 33.0 | 27/67 | 23/67 | 12/28 | 15/39 | 11/39 | 288 |
| pcdserver@zerocoder-0.8b-q4 | pcdServer | G | zerocoder-0.8b | fine-tuned | q4 | 0.53 | 46.4 | 27/67 | 24/67 | 12/28 | 15/39 | 12/39 | 270 |
| pcdserver@qwen2.5-0.5b-f16 | pcdServer | G | qwen2.5-0.5b | vanilla | f16 | 1.27 | 17.8 | 26/67 | 25/67 | 9/28 | 17/39 | 16/39 | 158 |
| pcdserver@qwen1.5-0.5b-q4 | pcdServer | G | qwen1.5-0.5b | vanilla | q4 | 0.41 | 21.6 | 26/67 | 28/67 | 8/28 | 18/39 | 21/39 | 217 |
| pcdserver@qwen3.5-0.8b-unsloth-iq2 | pcdServer | G | qwen3.5-0.8b-unsloth | vanilla | iq2 | 0.34 | 28.2 | 26/67 | 23/67 | 9/28 | 17/39 | 14/39 | 312 |
| pcdserver@qwen1.5-1.8b-q2 | pcdServer | G | qwen1.5-1.8b | vanilla | q2 | 0.86 | 39.8 | 26/67 | 36/67 | 10/28 | 16/39 | 26/39 | 392 |
| systemone@systemone-lite-0.5b † | SystemOne Lite (MPS) | G | systemone-lite-0.5b | dedicated | bf16/fp32 | 0.99 | 56.4 | 26/67 | 12/28 | 14/39 | |||
| ollaya@gliclass-large | ollaya | C | gliclass-large | dedicated | fp32 | 1.77 | 59.8 | 26/67 | 9/28 | 17/39 | 3,794 | ||
| pcdserver@qwen2.5-0.5b-q8 | pcdServer | G | qwen2.5-0.5b | vanilla | q8 | 0.68 | 17.8 | 25/67 | 24/67 | 8/28 | 17/39 | 16/39 | 40 |
| pcdserver@zerocoder-0.8b-q8 | pcdServer | G | zerocoder-0.8b | fine-tuned | q8 | 0.87 | 34.4 | 25/67 | 21/67 | 11/28 | 14/39 | 10/39 | 184 |
| pcdserver@zerocoder-0.8b-q6 | pcdServer | G | zerocoder-0.8b | fine-tuned | q6 | 0.68 | 35.3 | 24/67 | 21/67 | 11/28 | 13/39 | 10/39 | 182 |
| pcdserver@qwen2.5-0.5b-q6 | pcdServer | G | qwen2.5-0.5b | vanilla | q6 | 0.65 | 22.2 | 23/67 | 24/67 | 8/28 | 15/39 | 16/39 | 134 |
| pcdserver@qwen2.5-0.5b-q5 | pcdServer | G | qwen2.5-0.5b | vanilla | q5 | 0.52 | 18.9 | 22/67 | 22/67 | 7/28 | 15/39 | 15/39 | 219 |
| pcdserver@qwen1.5-0.5b-q2 | pcdServer | G | qwen1.5-0.5b | vanilla | q2 | 0.30 | 21.1 | 21/67 | 22/67 | 8/28 | 13/39 | 14/39 | 247 |
| pcdserver@qwen3-reranker-0.6b-q3 | pcdServer | G | qwen3-reranker-0.6b | fine-tuned | q3 | 0.35 | 24.2 | 21/67 | 21/67 | 8/28 | 13/39 | 13/39 | 89 |
| systemone@raz-nli-base † | Raz NLI (MPS) | G | raz-nli-base | dedicated | fp32 | 0.74 | 38.5 | 21/67 | 10/28 | 11/39 | |||
| dohnuts-metal@this-that-model-1.2-q2 | dohnuts (Metal) | G | this-that-model-1.2 | dedicated | q2 | 0.97 | 112.2 | 21/67 | 18/67 | 8/28 | 13/39 | 10/39 | |
| pcdserver@qwen2.5-0.5b-pcb-q3 | pcdServer | G | qwen2.5-0.5b-pcb | fine-tuned | q3 | 0.36 | 17.7 | 20/67 | 20/67 | 7/28 | 13/39 | 13/39 | 185 |
| ollaya@nli-modernbert-large | ollaya | G | nli-modernbert-large | dedicated | fp32 | 0.80 | 22.7 | 20/67 | 9/28 | 11/39 | 1,201 | ||
| pcdserver@euaiact-0.5b-q4 | pcdServer | G | euaiact-0.5b | fine-tuned | q4 | 0.40 | 23.8 | 20/67 | 20/67 | 8/28 | 12/39 | 12/39 | 697 |
| systemone@system-one-270m † | Transformers slot (MPS) | G | system-one-270m | dedicated | bf16/fp32 | 0.54 | 32.8 | 20/67 | 8/28 | 12/39 | |||
| pcdserver@qwen2.5-0.5b-pcb-f16 | pcdServer | G | qwen2.5-0.5b-pcb | fine-tuned | f16 | 0.99 | 17.2 | 19/67 | 20/67 | 7/28 | 12/39 | 13/39 | 144 |
| pcdserver@qwen2.5-0.5b-pcb-q6 | pcdServer | G | qwen2.5-0.5b-pcb | fine-tuned | q6 | 0.51 | 17.5 | 19/67 | 20/67 | 7/28 | 12/39 | 13/39 | 79 |
| pcdserver@qwen2.5-0.5b-pcb-q8 | pcdServer | G | qwen2.5-0.5b-pcb | fine-tuned | q8 | 0.53 | 17.5 | 19/67 | 20/67 | 7/28 | 12/39 | 13/39 | 40 |
| pcdserver@qwen2.5-0.5b-pcb-q5 | pcdServer | G | qwen2.5-0.5b-pcb | fine-tuned | q5 | 0.42 | 17.6 | 19/67 | 19/67 | 7/28 | 12/39 | 12/39 | 259 |
| pcdserver@qwen2.5-0.5b-q3 | pcdServer | G | qwen2.5-0.5b | vanilla | q3 | 0.43 | 18.4 | 18/67 | 19/67 | 5/28 | 13/39 | 14/39 | 316 |
| pcdserver@zerocoder-0.8b-q5 | pcdServer | G | zerocoder-0.8b | fine-tuned | q5 | 0.60 | 36.2 | 18/67 | 17/67 | 8/28 | 10/39 | 9/39 | 216 |
| dohnuts-metal@decider-0.8b-q2 | dohnuts (Metal) | G | decider-0.8b | dedicated | q2 | 0.42 | 55.2 | 18/67 | 17/67 | 7/28 | 11/39 | 10/39 | |
| slot@decider-0.8b-q2 | slot (llama-server) | G | decider-0.8b | dedicated | q2 | 0.42 | 64.7 | 18/67 | 17/67 | 7/28 | 11/39 | 10/39 | 108 |
| pcdserver@qwen2.5-0.5b-pcb-q4 | pcdServer | G | qwen2.5-0.5b-pcb | fine-tuned | q4 | 0.40 | 22.9 | 17/67 | 19/67 | 6/28 | 11/39 | 13/39 | 361 |
| pcdserver@decider-0.8b-q2 | pcdServer | G | decider-0.8b | dedicated | q2 | 0.42 | 29.3 | 17/67 | 17/67 | 4/28 | 13/39 | 13/39 | 577 |
| systemone@oomu-systemone-0.6b † | llama-cpp-python (Metal) | G | oomu-systemone-0.6b | dedicated | q8_0 | 0.64 | 65.2 | 17/67 | 17/67 | 7/28 | 10/39 | 10/39 | |
| pcdserver@qwen2.5-0.5b-q2 | pcdServer | G | qwen2.5-0.5b | vanilla | q2 | 0.42 | 17.6 | 15/67 | 16/67 | 4/28 | 11/39 | 12/39 | 477 |
| pcdserver@qwen3-4b-unsloth-iq1 | pcdServer | G | qwen3-4b-unsloth | vanilla | iq1 | 1.08 | 72.5 | 13/67 | 14/67 | 6/28 | 7/39 | 8/39 | 594 |
| pcdserver@lq-decide-0.6b-q8 | pcdServer | G | lq-decide-0.6b | fine-tuned | q8 | 0.64 | 23.1 | 11/67 | 11/67 | 4/28 | 7/39 | 7/39 | 135 |
| pcdserver@qwen3-reranker-0.6b-q2 | pcdServer | G | qwen3-reranker-0.6b | fine-tuned | q2 | 0.30 | 23.9 | 10/67 | 14/67 | 5/28 | 5/39 | 9/39 | 241 |
| pcdserver@lq-decide-0.6b-q4 | pcdServer | G | lq-decide-0.6b | fine-tuned | q4 | 0.40 | 29.5 | 10/67 | 11/67 | 4/28 | 6/39 | 7/39 | 183 |
CPU and mixed runs¶
The audit found 19 CPU-only runs and 19 runs with an explicit CPU/GPU split. This table identifies them all. It excludes CPU/Neural Engine and automatic-placement runs; those have their own codes in the full table. Historical labels describe recorded configurations, rather than a per-operation utilization trace.
| Method | C/G | Execution | Evidence | ms/query | Translated | Direct |
|---|---|---|---|---|---|---|
| pcdserver@rune-mradermacher-26b-a4b-q5 † | C+G | CPU weights (--gpu-layers 0); Metal computation and KV offload enabled | configured | 962.3 | 65/67 | 65/67 |
| pcdserver@rune-mradermacher-26b-a4b-q8 † | C | CPU only (--gpu-layers 0 --cpu); compute and KV offload disabled | configured | 758.1 | 64/67 | 64/67 |
| gguf-head@jev-omni-q3 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 1,222.3 | 64/67 | 64/67 |
| pcdserver@rune-mradermacher-26b-a4b-q4 † | C+G | CPU weights (--gpu-layers 0); Metal computation and KV offload enabled | configured | 487.1 | 63/67 | 64/67 |
| gguf-head@jev-omni-q8 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 1,124.1 | 63/67 | 63/67 |
| gguf-head@jev-omni-q5 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 1,212.7 | 63/67 | 63/67 |
| dohnuts@decider-0.8b-dreamblooms-q8 | C | CPU; recorded runtime/launch configuration | configured | 275.2 | 62/67 | 60/67 |
| coreml@gliner25-decide-fp16-l256-cpu-gpu † | C+G | Core ML explicitly CPU_AND_GPU; Neural Engine disabled; exact operation placement unprofiled | configured | 21.9 | 61/67 | |
| onnx@gliner25-decide-fp32 | C | ONNX Runtime CPUExecutionProvider | configured | 67.3 | 61/67 | |
| ollaya@decider-0.8b | C | ollaya ONNX Runtime CPU on this Mac; recorded runtime recipe | configured | 284.8 | 61/67 | 60/67 |
| coreml@kev-0.8b-fp16-l512 † | C+G | Core ML CPU_AND_GPU; exact operation placement unprofiled | configured | 121.3 | 60/67 | 62/67 |
| ollaya@kev-0.8b | C | ollaya ONNX Runtime CPU on this Mac; recorded runtime recipe | configured | 268.2 | 60/67 | 62/67 |
| ollaya@von-1.1 | C | ollaya ONNX Runtime CPU on this Mac; recorded runtime recipe | configured | 62.8 | 57/67 | |
| onnx@von-1.2 | C | ONNX Runtime CPUExecutionProvider | configured | 52.6 | 57/67 | |
| ollaya@decision-eos | C | ollaya ONNX Runtime CPU on this Mac; recorded runtime recipe | configured | 322.9 | 57/67 | 56/67 |
| gguf-head@laya-neutron-ml | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 42.7 | 54/67 | 51/67 |
| systemone@system-one-scorer-onnx † | C | ONNX Runtime CPUExecutionProvider; scalar Q4 graph, CPU tokenization | configured | 1,410.8 | 53/67 | |
| laya-ml-onnx-fp32 | C | ONNX Runtime CPUExecutionProvider | configured | 23.4 | 52/67 | 49/67 |
| onnx@laya-ml-fp32 | C | ONNX Runtime CPUExecutionProvider | configured | 24.7 | 52/67 | 49/67 |
| laya-ml-onnx-int8 | C | ONNX Runtime CPUExecutionProvider | configured | 25.5 | 52/67 | 49/67 |
| laya-wd-gguf-q8 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 44.4 | 52/67 | 49/67 |
| laya-wd-gguf-f16 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 44.4 | 52/67 | 49/67 |
| laya-ml-gguf | C | CPU; recorded runtime/launch configuration | configured | 57.6 | 52/67 | 49/67 |
| onnx@laya-ml-fp16 | C | ONNX Runtime CPUExecutionProvider | configured | 57.6 | 52/67 | 49/67 |
| laya-ml-gguf-f16 | C | CPU; recorded runtime/launch configuration | configured | 61.4 | 52/67 | 49/67 |
| laya-wd-gguf-q6 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 44.5 | 51/67 | 47/67 |
| laya-ml-gguf-q4 | C | CPU; recorded runtime/launch configuration | configured | 55.0 | 51/67 | 47/67 |
| laya-wd-gguf-q3 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 45.3 | 49/67 | 43/67 |
| laya-wd-gguf-q5 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 45.5 | 47/67 | 43/67 |
| gguf-head@laya-neutron-en | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 62.4 | 46/67 | |
| systemone@clm-8b-q8 | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 107.6 | 39/67 | 28/67 |
| systemone@clm-8b-mlx6 † | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 415.7 | 39/67 | |
| ollaya@clm-8b | C | ollaya ONNX Runtime CPU on this Mac; recorded runtime recipe | configured | 1,677.9 | 39/67 | 30/67 |
| systemone@clm-8b-mlx8 | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 100.4 | 38/67 | 31/67 |
| systemone@clm-8b-q5 | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 119.2 | 38/67 | 27/67 |
| systemone@clm-8b-q6-k † | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 490.7 | 38/67 | |
| systemone@clm-8b-q5-km † | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 509.9 | 38/67 | |
| onnx@decima-small-int8 | C | ONNX Runtime CPUExecutionProvider | configured | 12.4 | 37/67 | 38/67 |
| laya-wd-gguf-q4 | C+G | Metal encoder; CPU decision head (NumPy/PyTorch) | configured | 43.3 | 36/67 | 38/67 |
| systemone@clm-8b-q4 | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 113.8 | 36/67 | 28/67 |
| systemone@clm-8b-q4-km † | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 449.4 | 36/67 | |
| systemone@clm-8b-mlx4 | C+G | Metal/MLX backbone; CLM decision head configured --device cpu | configured | 89.6 | 33/67 | 28/67 |
| onnx@julia-1 | C | ONNX Runtime CPUExecutionProvider | configured | 12.2 | 31/67 | 31/67 |
| ollaya@gliclass-large | C | ollaya ONNX Runtime CPU on this Mac; recorded runtime recipe | configured | 59.8 | 26/67 |
One set of weights on several engines¶
An engine is not a neutral container. This table puts the same model file side by side on every engine that ran it. decider-0.8b scores 61 or 62 through its own trained readout on dohnuts and llama-server, but 55 on pcdServer, which reads it as a chat model. Chapter 6 explains the difference.
| Model | Quant | Engine | C/G | ms/query | Translated | Direct |
|---|---|---|---|---|---|---|
| decider-0.8b | f16 | pcdServer | G | 30.7 | 55/67 | 56/67 |
| decider-0.8b | f16 | dohnuts (Metal) | G | 51.9 | 61/67 | 60/67 |
| decider-0.8b | f16 | slot (llama-server) | G | 63.6 | 61/67 | 60/67 |
| decider-0.8b | fp16 | ExecuTorch (MLX) | G | 89.7 | 61/67 | 60/67 |
| decider-0.8b | fp32 | ollaya | C | 284.8 | 61/67 | 60/67 |
| decider-0.8b | q2 | pcdServer | G | 29.3 | 17/67 | 17/67 |
| decider-0.8b | q2 | dohnuts (Metal) | G | 55.2 | 18/67 | 17/67 |
| decider-0.8b | q2 | slot (llama-server) | G | 64.7 | 18/67 | 17/67 |
| decider-0.8b | q3 | pcdServer | G | 29.3 | 52/67 | 47/67 |
| decider-0.8b | q3 | dohnuts (Metal) | G | 54.5 | 57/67 | 55/67 |
| decider-0.8b | q3 | slot (llama-server) | G | 69.7 | 57/67 | 55/67 |
| decider-0.8b | q4 | pcdServer | G | 40.2 | 57/67 | 57/67 |
| decider-0.8b | q4 | dohnuts (Metal) | G | 53.3 | 61/67 | 59/67 |
| decider-0.8b | q4 | slot (llama-server) | G | 67.4 | 61/67 | 59/67 |
| decider-0.8b | q5 | pcdServer | G | 29.9 | 57/67 | 57/67 |
| decider-0.8b | q5 | dohnuts (Metal) | G | 55.2 | 62/67 | 60/67 |
| decider-0.8b | q5 | slot (llama-server) | G | 67.9 | 62/67 | 60/67 |
| decider-0.8b | q6 | pcdServer | G | 29.8 | 55/67 | 57/67 |
| decider-0.8b | q6 | dohnuts (Metal) | G | 53.5 | 62/67 | 62/67 |
| decider-0.8b | q6 | slot (llama-server) | G | 68.1 | 62/67 | 62/67 |
| decider-0.8b | q8 | pcdServer | G | 30.2 | 55/67 | 57/67 |
| decider-0.8b | q8 | dohnuts (Metal) | G | 51.4 | 62/67 | 61/67 |
| decider-0.8b | q8 | slot (llama-server) | G | 64.0 | 62/67 | 61/67 |
| decider-0.8b-dreamblooms | q8 | pcdServer | G | 28.3 | 55/67 | 57/67 |
| decider-0.8b-dreamblooms | q8 | dohnuts (Metal) | G | 52.2 | 62/67 | 61/67 |
| decider-0.8b-dreamblooms | q8 | slot (llama-server) | G | 65.1 | 62/67 | 61/67 |
| decider-0.8b-dreamblooms | q8 | dohnuts (CPU) | C | 275.2 | 62/67 | 60/67 |
| decider-2b | f16 | pcdServer | G | 42.2 | 60/67 | 60/67 |
| decider-2b | f16 | dohnuts (Metal) | G | 100.3 | 59/67 | 62/67 |
| decider-2b | f16 | slot (llama-server) | G | 112.4 | 59/67 | 62/67 |
| decider-2b | fp16 | ExecuTorch (MLX) | G | 251.3 | 59/67 | 62/67 |
| decider-2b | iq4 | pcdServer | G | 41.4 | 57/67 | 57/67 |
| decider-2b | iq4 | dohnuts (Metal) | G | 106.7 | 56/67 | 57/67 |
| decider-2b | iq4 | slot (llama-server) | G | 126.8 | 56/67 | 57/67 |
| decider-2b-dreamblooms | q8 | pcdServer | G | 40.1 | 59/67 | 60/67 |
| decider-2b-dreamblooms | q8 | dohnuts (Metal) | G | 103.7 | 59/67 | 62/67 |
| decider-2b-dreamblooms | q8 | slot (llama-server) | G | 114.4 | 59/67 | 62/67 |
| decider-35b-a3b | q4 | dohnuts (Metal) | G | 266.3 | 64/67 | 64/67 |
| decider-35b-a3b | q4 | pcdServer | G | 296.8 | 63/67 | 63/67 |
| decider-35b-a3b | q4 | slot (llama-server) | G | 299.1 | 64/67 | 64/67 |
| decider-4b | bf16 | PyTorch | G | 348.2 | 63/67 | 65/67 |
Best model per family¶
The best row of every model, grouped by family. A dedicated model was trained for this kind of question and has its own readout; a fine-tuned model is a chat model tuned by someone for a task; a vanilla model is a stock chat model. Best method names the row the other columns come from. Chapter 4 describes the families.
| Family | Model | Engine | C/G | GB | ms/query | Translated | Direct | Best method |
|---|---|---|---|---|---|---|---|---|
| dedicated | typesafe/jev-1.13 | OpenRouter API | R | 364.4 | 65/67 | 65/67 | openrouter@typesafe/jev-1.13 | |
| dedicated | rune-26b-a4b | slot (llama-server) | G | 13.52 | 366.6 | 65/67 | 57/67 | slot@rune-26b-a4b-q3 |
| dedicated | rune-mradermacher-26b-a4b | pcdServer | C+G | 19.13 | 962.3 | 65/67 | 65/67 | pcdserver@rune-mradermacher-26b-a4b-q5 |
| dedicated | jaredpalmer/kev-4b | OpenRouter API | R | 1,653.0 | 65/67 | 65/67 | openrouter@jaredpalmer/kev-4b | |
| dedicated | decider-35b-a3b | dohnuts (Metal) | G | 21.17 | 266.3 | 64/67 | 64/67 | dohnuts-metal@decider-35b-a3b-q4 |
| dedicated | neohorse-4b | System One server | G | 5.17 | 273.1 | 64/67 | 64/67 | systemone@neohorse-4b-q8 |
| dedicated | jpt-4b | dohnuts (Metal) | G | 4.48 | 335.6 | 64/67 | 64/67 | dohnuts-metal@jpt-4b-q8 |
| dedicated | togethercomputer/tev1-4b-experimental | OpenRouter API | R | 339.3 | 64/67 | 65/67 | openrouter@togethercomputer/tev1-4b-experimental | |
| dedicated | inception/mercury-decide:free | OpenRouter API | R | 408.3 | 64/67 | 64/67 | openrouter@inception/mercury-decide:free | |
| dedicated | liquid/d1 | OpenRouter API | R | 566.2 | 64/67 | 64/67 | openrouter@liquid/d1 | |
| dedicated | winnow-12b | ollaya | G | 12.67 | 891.1 | 64/67 | 65/67 | ollaya@winnow-12b |
| dedicated | jev-omni | llama.cpp + MLX head | G | 6.09 | 908.4 | 64/67 | 64/67 | gguf-head@jev-omni-q3-last-mlx |
| dedicated | jpt-9b | pcdServer | G | 6.47 | 155.2 | 63/67 | 64/67 | pcdserver@jpt-9b-q5 |
| dedicated | openjev-35b-a3b | MLX | G | 20.99 | 250.0 | 63/67 | 62/67 | mlx@openjev-35b-a3b-4bit |
| dedicated | jeb-35b-a3b | System One server | G | 17.17 | 333.2 | 63/67 | 63/67 | systemone@jeb-35b-a3b-q3 |
| dedicated | decider-4b | PyTorch | G | 348.2 | 63/67 | 65/67 | torch@decider-4b-bf16 | |
| dedicated | clef-flash | Clef joint head (MLX) | G | 5.28 | 554.8 | 63/67 | 63/67 | systemone@clef-flash-mlx4 |
| dedicated | lev-4b | System One server | G | 9.49 | 2,203.8 | 63/67 | 64/67 | systemone@lev-4b |
| dedicated | decider-0.8b | dohnuts (Metal) | G | 0.81 | 51.4 | 62/67 | 61/67 | dohnuts-metal@decider-0.8b-q8 |
| dedicated | decider-0.8b-dreamblooms | dohnuts (Metal) | G | 0.81 | 52.2 | 62/67 | 61/67 | dohnuts-metal@decider-0.8b-dreamblooms-q8 |
| dedicated | system-one-minicpm5-2b-q8 | System One (MLX) | G | 2.67 | 141.8 | 62/67 | 59/67 | systemone@system-one-minicpm5-2b-q8 |
| dedicated | jet-4b | dohnuts (Metal) | G | 4.48 | 281.4 | 62/67 | 63/67 | dohnuts-metal@jet-4b-q8 |
| dedicated | jevk5-4b | slot (llama-server) | G | 2.71 | 350.5 | 62/67 | 63/67 | slot@jevk5-4b-q4 |
| dedicated | imajev-4b | System One server | G | 9.81 | 363.1 | 62/67 | 64/67 | systemone@imajev-4b |
| dedicated | kev-4b | laya.cpp | G | 1,103.1 | 62/67 | 55/67 | kev-4b-gguf | |
| dedicated | upstage/solar-decide | OpenRouter API | R | 1,253.3 | 62/67 | 63/67 | openrouter@upstage/solar-decide | |
| dedicated | gliner25-decide | Core ML | C+G | 0.93 | 21.9 | 61/67 | coreml@gliner25-decide-fp16-l256-cpu-gpu | |
| dedicated | decider-2b | pcdServer | G | 3.78 | 42.2 | 60/67 | 60/67 | pcdserver@decider-2b-f16 |
| dedicated | tev1-0.8b | MLX | G | 0.42 | 64.8 | 60/67 | 59/67 | mlx@tev1-0.8b-4bit |
| dedicated | kev-0.8b | Core ML | C+G | 1.51 | 121.3 | 60/67 | 62/67 | coreml@kev-0.8b-fp16-l512 |
| dedicated | decider-2b-dreamblooms | pcdServer | G | 2.01 | 40.1 | 59/67 | 60/67 | pcdserver@decider-2b-dreamblooms-q8 |
| dedicated | von-1.2 | ONNX Runtime | C | 1.58 | 52.6 | 57/67 | onnx@von-1.2 | |
| dedicated | von-1.1 | ollaya | C | 1.59 | 62.8 | 57/67 | ollaya@von-1.1 | |
| dedicated | this-that-model-1.2 | dohnuts (Metal) | G | 1.10 | 113.1 | 57/67 | 56/67 | dohnuts-metal@this-that-model-1.2-q3 |
| dedicated | leo-1.7b | System One server | G | 3.52 | 136.2 | 57/67 | 62/67 | systemone@leo-1.7b |
| dedicated | decision-eos | ollaya | C | 1.54 | 322.9 | 57/67 | 56/67 | ollaya@decision-eos |
| dedicated | respan/span-01-lite:free | OpenRouter API | R | 436.0 | 56/67 | 60/67 | openrouter@respan/span-01-lite:free | |
| dedicated | respan/span-01-lite | OpenRouter API | R | 447.5 | 56/67 | 60/67 | openrouter@respan/span-01-lite | |
| dedicated | respan/span-01 | OpenRouter API | R | 487.4 | 56/67 | 60/67 | openrouter@respan/span-01 | |
| dedicated | bosun-1.7b | llama-cpp-python (Metal) | G | 1.26 | 130.8 | 55/67 | 55/67 | systemone@bosun-1.7b-q5 |
| dedicated | laya-multilingual | llama.cpp embeddings + head | C+G | 0.27 | 42.7 | 54/67 | 51/67 | gguf-head@laya-neutron-ml |
| dedicated | system-one-scorer-onnx | ONNX scalar (CPU) | C | 2.64 | 1,410.8 | 53/67 | systemone@system-one-scorer-onnx | |
| dedicated | semif-4b-nli | MLX | G | 8.41 | 120.6 | 51/67 | 49/67 | mlx@semif-4b-nli |
| dedicated | system-one-qwen3-0.6b | System One (MLX) | G | 1.52 | 47.7 | 49/67 | 52/67 | systemone@system-one-qwen3-0.6b |
| dedicated | laya-typed-decisions | PyTorch | G | 17.7 | 48/67 | 43/67 | laya-typed-decisions | |
| dedicated | laya-en | llama.cpp embeddings + head | C+G | 0.29 | 62.4 | 46/67 | gguf-head@laya-neutron-en | |
| dedicated | xDecision | xDecision (MLX) | G | 0.40 | 14.3 | 45/67 | 43/67 | systemone@xdecision-q8-mlx |
| dedicated | system-one-distilled | Transformers NLI (MPS) | G | 0.28 | 21.2 | 42/67 | systemone@system-one-distilled | |
| dedicated | system-one-gold | Transformers NLI (MPS) | G | 0.28 | 21.4 | 42/67 | systemone@system-one-gold | |
| dedicated | raz-nli-xsmall-openjev | Raz NLI (MPS) | G | 0.28 | 49.4 | 42/67 | systemone@raz-nli-xsmall-openjev | |
| dedicated | clm-8b | System One server | C+G | 8.25 | 107.6 | 39/67 | 28/67 | systemone@clm-8b-q8 |
| dedicated | decima-small | ONNX Runtime | C | 0.13 | 12.4 | 37/67 | 38/67 | onnx@decima-small-int8 |
| dedicated | system-one-zeroshot | Transformers NLI (MPS) | G | 0.28 | 21.2 | 37/67 | systemone@system-one-zeroshot | |
| dedicated | pulse-decide-150m | PyTorch | G | 0.60 | 17.9 | 34/67 | torch@pulse-decide-150m | |
| dedicated | lumma-fev-0.6b | PyTorch | G | 1.30 | 69.5 | 33/67 | 28/67 | torch@lumma-fev-0.6b |
| dedicated | julia-1 | MLX | G | 0.58 | 6.8 | 31/67 | 31/67 | mlx@julia-1 |
| dedicated | systemone-lite-0.5b | SystemOne Lite (MPS) | G | 0.99 | 56.4 | 26/67 | systemone@systemone-lite-0.5b | |
| dedicated | gliclass-large | ollaya | C | 1.77 | 59.8 | 26/67 | ollaya@gliclass-large | |
| dedicated | raz-nli-base | Raz NLI (MPS) | G | 0.74 | 38.5 | 21/67 | systemone@raz-nli-base | |
| dedicated | nli-modernbert-large | ollaya | G | 0.80 | 22.7 | 20/67 | ollaya@nli-modernbert-large | |
| dedicated | system-one-270m | Transformers slot (MPS) | G | 0.54 | 32.8 | 20/67 | systemone@system-one-270m | |
| dedicated | oomu-systemone-0.6b | llama-cpp-python (Metal) | G | 0.64 | 65.2 | 17/67 | 17/67 | systemone@oomu-systemone-0.6b |
| fine-tuned | qwen3.5-4b-hmm | pcdServer | G | 4.48 | 87.6 | 64/67 | 65/67 | pcdserver@qwen3.5-4b-hmm-q8 |
| fine-tuned | qwen3.8-4b-distill | pcdServer | G | 2.78 | 93.6 | 62/67 | 63/67 | pcdserver@qwen3.8-4b-distill-q4 |
| fine-tuned | qwen3.8-9b-distill | pcdServer | G | 5.78 | 137.8 | 62/67 | 64/67 | pcdserver@qwen3.8-9b-distill-q4 |
| fine-tuned | ornith-1.5-35b-a3b-splash | Splash (Metal) | G | 20.95 | 322.4 | 62/67 | 63/67 | splash@ornith-1.5-35b-a3b-splash |
| fine-tuned | qwen3.5-4b-supercoder | pcdServer | G | 2.61 | 84.2 | 61/67 | 62/67 | pcdserver@qwen3.5-4b-supercoder-q4 |
| fine-tuned | qwen2.5-coder-7b | pcdServer | G | 4.68 | 108.9 | 61/67 | 62/67 | pcdserver@qwen2.5-coder-7b-q4 |
| fine-tuned | qwen2.5-coder-3b | pcdServer | G | 2.10 | 60.2 | 60/67 | 59/67 | pcdserver@qwen2.5-coder-3b-q4 |
| fine-tuned | pulse-0.6b | pcdServer | G | 0.40 | 27.9 | 54/67 | 52/67 | pcdserver@pulse-0.6b-q4 |
| fine-tuned | lq-decide-1.7b | pcdServer | G | 1.83 | 35.1 | 50/67 | 46/67 | pcdserver@lq-decide-1.7b-q8 |
| fine-tuned | qwen3.8-2b-distill | pcdServer | G | 1.31 | 42.9 | 49/67 | 51/67 | pcdserver@qwen3.8-2b-distill-q4 |
| fine-tuned | zerocoder-0.8b | pcdServer | G | 0.38 | 34.7 | 42/67 | 36/67 | pcdserver@zerocoder-0.8b-q2 |
| fine-tuned | qwen3-reranker-0.6b | pcdServer | G | 0.44 | 24.7 | 40/67 | 44/67 | pcdserver@qwen3-reranker-0.6b-q5 |
| fine-tuned | lyco-router-0.6b | pcdServer | G | 0.48 | 26.4 | 35/67 | 33/67 | pcdserver@lyco-router-0.6b-q4 |
| fine-tuned | qwen2.5-0.5b-pcb | pcdServer | G | 0.34 | 18.1 | 27/67 | 23/67 | pcdserver@qwen2.5-0.5b-pcb-q2 |
| fine-tuned | euaiact-0.5b | pcdServer | G | 0.40 | 23.8 | 20/67 | 20/67 | pcdserver@euaiact-0.5b-q4 |
| fine-tuned | lq-decide-0.6b | pcdServer | G | 0.64 | 23.1 | 11/67 | 11/67 | pcdserver@lq-decide-0.6b-q8 |
| vanilla | qwen3.5-4b-unsloth | pcdServer | G | 4.48 | 85.0 | 63/67 | 64/67 | pcdserver@qwen3.5-4b-unsloth-q8 |
| vanilla | qwen2.5-3b | pcdServer | G | 3.29 | 53.3 | 62/67 | 57/67 | pcdserver@qwen2.5-3b-q8 |
| vanilla | qwen3.5-4b | pcdServer | G | 3.01 | 89.8 | 62/67 | 63/67 | pcdserver@qwen3.5-4b-q4 |
| vanilla | bonsai-27b | pcdServer | G | 3.80 | 513.5 | 62/67 | 63/67 | pcdserver@bonsai-27b-q1 |
| vanilla | qwen3.5-2b | pcdServer | G | 1.22 | 42.7 | 61/67 | 62/67 | pcdserver@qwen3.5-2b-q3 |
| vanilla | qwen3-4b-unsloth | pcdServer | G | 8.05 | 71.5 | 61/67 | 62/67 | pcdserver@qwen3-4b-unsloth-bf16 |
| vanilla | qwen3-4b | pcdServer | G | 2.50 | 75.5 | 61/67 | 62/67 | pcdserver@qwen3-4b-q4 |
| vanilla | qwen3-8b | pcdServer | G | 5.03 | 116.0 | 61/67 | 62/67 | pcdserver@qwen3-8b-q4 |
| vanilla | qwen3.5-9b-unsloth | pcdServer | G | 5.68 | 136.9 | 61/67 | 63/67 | pcdserver@qwen3.5-9b-unsloth-q4 |
| vanilla | qwen3.5-0.8b | pcdServer | G | 0.65 | 32.1 | 57/67 | 57/67 | pcdserver@qwen3.5-0.8b-q5 |
| vanilla | qwen3-1.7b-unsloth | pcdServer | G | 1.83 | 35.1 | 57/67 | 59/67 | pcdserver@qwen3-1.7b-unsloth-q8 |
| vanilla | qwen3-1.7b | pcdServer | G | 2.17 | 35.2 | 57/67 | 59/67 | pcdserver@qwen3-1.7b-q8 |
| vanilla | qwen3.5-2b-unsloth | pcdServer | G | 2.01 | 39.5 | 57/67 | 59/67 | pcdserver@qwen3.5-2b-unsloth-q8 |
| vanilla | qwen3.5-0.8b-unsloth | pcdServer | G | 0.59 | 29.6 | 56/67 | 57/67 | pcdserver@qwen3.5-0.8b-unsloth-q5 |
| vanilla | qwen3.5-0.8b-ggml | pcdServer | G | 1.56 | 30.3 | 55/67 | 57/67 | pcdserver@qwen3.5-0.8b-ggml-bf16 |
| vanilla | bonsai-1.7b | pcdServer | G | 0.49 | 34.9 | 54/67 | 55/67 | pcdserver@bonsai-1.7b-q2 |
| vanilla | qwen3-0.6b | pcdServer | G | 0.64 | 25.5 | 53/67 | 54/67 | pcdserver@qwen3-0.6b-q8 |
| vanilla | qwen1.5-4b | pcdServer | G | 4.20 | 247.8 | 42/67 | 43/67 | pcdserver@qwen1.5-4b-q8 |
| vanilla | qwen1.5-0.5b | pcdServer | G | 0.35 | 21.2 | 35/67 | 32/67 | pcdserver@qwen1.5-0.5b-q3 |
| vanilla | qwen1.5-1.8b | pcdServer | G | 1.02 | 41.8 | 34/67 | 35/67 | pcdserver@qwen1.5-1.8b-q3 |
| vanilla | qwen2.5-0.5b | pcdServer | G | 0.49 | 20.2 | 28/67 | 26/67 | pcdserver@qwen2.5-0.5b-q4 |
Quantization sweeps¶
Each row is one model on one engine, measured at several quantizations. Each cell is the translated score out of 67, then the mean time per query; an empty cell was not measured. Between Q4 and Q8 the scores barely move, and below Q3 they collapse. Chapter 7 reads the grid.
| Model | Engine | C/G | iq1 | q2 | iq2 | q3 | iq3 | q4 | mlx4 | q5 | q6 | q8 | mlx8 | f16 | bf16 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| clef-flash | llama-server (System One) | G | |||||||||||||
| clm-8b | System One server | C+G | 36 · 114 ms | 33 · 90 ms | 38 · 119 ms | 39 · 108 ms | 38 · 100 ms | ||||||||
| decider-0.8b | dohnuts (Metal) | G | 18 · 55 ms | 57 · 54 ms | 61 · 53 ms | 62 · 55 ms | 62 · 54 ms | 62 · 51 ms | 61 · 52 ms | ||||||
| decider-0.8b | pcdServer | G | 17 · 29 ms | 52 · 29 ms | 57 · 40 ms | 57 · 30 ms | 55 · 30 ms | 55 · 30 ms | 55 · 31 ms | ||||||
| decider-0.8b | slot (llama-server) | G | 18 · 65 ms | 57 · 70 ms | 61 · 67 ms | 62 · 68 ms | 62 · 68 ms | 62 · 64 ms | 61 · 64 ms | ||||||
| jpt-4b | dohnuts (Metal) | G | 61 · 315 ms | 62 · 291 ms | 63 · 367 ms | 64 · 336 ms | |||||||||
| jpt-4b | pcdServer | G | 60 · 94 ms | 63 · 93 ms | 62 · 99 ms | 63 · 86 ms | |||||||||
| jpt-9b | dohnuts (Metal) | G | 61 · 637 ms | 63 · 620 ms | 63 · 620 ms | 63 · 625 ms | |||||||||
| jpt-9b | pcdServer | G | 60 · 139 ms | 61 · 137 ms | 63 · 155 ms | 61 · 142 ms | |||||||||
| laya-multilingual | llama.cpp embeddings | C+G | 49 · 45 ms | 36 · 43 ms | 47 · 45 ms | 51 · 45 ms | 52 · 44 ms | 52 · 44 ms | |||||||
| qwen1.5-0.5b | pcdServer | G | 21 · 21 ms | 35 · 21 ms | 26 · 22 ms | 28 · 22 ms | 30 · 23 ms | 28 · 21 ms | |||||||
| qwen1.5-1.8b | pcdServer | G | 26 · 40 ms | 34 · 42 ms | 33 · 40 ms | 32 · 41 ms | 31 · 40 ms | 32 · 36 ms | |||||||
| qwen1.5-4b | pcdServer | G | 41 · 233 ms | 39 · 237 ms | 41 · 232 ms | 39 · 284 ms | 40 · 276 ms | 42 · 248 ms | |||||||
| qwen2.5-0.5b | pcdServer | G | 15 · 18 ms | 18 · 18 ms | 28 · 20 ms | 22 · 19 ms | 23 · 22 ms | 25 · 18 ms | 26 · 18 ms | ||||||
| qwen2.5-0.5b-pcb | pcdServer | G | 27 · 18 ms | 20 · 18 ms | 17 · 23 ms | 19 · 18 ms | 19 · 17 ms | 19 · 17 ms | 19 · 17 ms | ||||||
| qwen2.5-3b | pcdServer | G | 53 · 58 ms | 49 · 56 ms | 57 · 61 ms | 62 · 59 ms | 61 · 62 ms | 62 · 62 ms | 62 · 53 ms | 62 · 54 ms | |||||
| qwen3-1.7b | pcdServer | G | 54 · 38 ms | 50 · 38 ms | 54 · 41 ms | 57 · 39 ms | 57 · 39 ms | 57 · 35 ms | 57 · 36 ms | ||||||
| qwen3-1.7b-unsloth | pcdServer | G | 44 · 37 ms | 53 · 38 ms | 56 · 40 ms | 57 · 39 ms | 57 · 39 ms | 57 · 35 ms | |||||||
| qwen3-4b-unsloth | pcdServer | G | 13 · 72 ms | 59 · 73 ms | 60 · 78 ms | 61 · 76 ms | 60 · 79 ms | 60 · 79 ms | 60 · 70 ms | 61 · 72 ms | |||||
| qwen3-reranker-0.6b | pcdServer | G | 10 · 24 ms | 21 · 24 ms | 28 · 29 ms | 40 · 25 ms | 33 · 25 ms | 30 · 23 ms | 30 · 24 ms | ||||||
| qwen3.5-0.8b | pcdServer | G | 40 · 32 ms | 46 · 31 ms | 52 · 30 ms | 57 · 32 ms | 55 · 29 ms | 55 · 31 ms | 55 · 30 ms | ||||||
| qwen3.5-0.8b-unsloth | pcdServer | G | 43 · 28 ms | 26 · 28 ms | 48 · 30 ms | 34 · 28 ms | 55 · 50 ms | 56 · 30 ms | 55 · 29 ms | 55 · 29 ms | 55 · 31 ms | ||||
| qwen3.5-2b | pcdServer | G | 49 · 41 ms | 61 · 43 ms | 57 · 47 ms | 57 · 42 ms | 57 · 43 ms | 57 · 39 ms | |||||||
| qwen3.5-2b-unsloth | pcdServer | G | 53 · 41 ms | 44 · 41 ms | 57 · 42 ms | 57 · 40 ms | 57 · 45 ms | 57 · 43 ms | 56 · 44 ms | 57 · 40 ms | 57 · 42 ms | ||||
| qwen3.5-4b-unsloth | pcdServer | G | 61 · 94 ms | 52 · 87 ms | 60 · 93 ms | 61 · 90 ms | 61 · 94 ms | 63 · 94 ms | 62 · 94 ms | 63 · 85 ms | |||||
| qwen3.8-2b-distill | pcdServer | G | 49 · 43 ms | 45 · 46 ms | 43 · 43 ms | 44 · 39 ms | |||||||||
| this-that-model-1.2 | dohnuts (Metal) | G | 21 · 112 ms | 57 · 113 ms | 51 · 110 ms | 54 · 117 ms | 53 · 108 ms | ||||||||
| zerocoder-0.8b | pcdServer | G | 42 · 35 ms | 32 · 35 ms | 27 · 46 ms | 18 · 36 ms | 24 · 35 ms | 25 · 34 ms | 27 · 33 ms |
Every query¶
Every one of the 67 queries, hardest first. Right, translated and right, direct are the share of all methods that answered it correctly in each mode. jev is jev's answer on the translated text, which you can compare with the labelled task. Chapter 9 discusses the three hardest.
| Query | Lang | Task | Right, translated | Right, direct | jev |
|---|---|---|---|---|---|
| How do I write a liga feature? | en | docs | 12 % | 12 % | fea |
| Which words should I look at to judge the spacing of round letters? | en | sample | 23 % | 23 % | sample |
| Show me how a glyph is stored in a VFJ file | en | docs | 26 % | 26 % | vfj |
| Text with lots of kerning pairs | en | sample | 38 % | 38 % | sample |
| VFJ 형식으로 이름이 hyphen이고 너비가 300인 글리프 정의를 작성해 줘 | ko | vfj | 39 % | 90 % | vfj |
| Loop through masters and print their names | en | python | 48 % | 48 % | python |
| Rename all glyphs in every open font, I don't want to click | en | python | 53 % | 53 % | python |
| Positioning rule that lifts the acute over capital letters | en | fea | 57 % | 57 % | fea |
| Can FontLab batch rename glyphs? | en | docs | 60 % | 60 % | docs |
| Script the kerning for A V | en | fea | 60 % | 60 % | fea |
| I need substitutions so that 1/2 becomes onehalf | en | fea | 61 % | 61 % | fea |
| پائتھن اسکرپٹ لکھو جو تمام گلفس کی چوڑائی 600 کر دے | ur | python | 62 % | 69 % | python |
| Can you automate adding sidebearings of 40 to every glyph in the selection? | en | python | 64 % | 64 % | python |
| Donne-moi un pangramme français pour tester mes accents | fr | sample | 65 % | 67 % | sample |
| תן לי טקסט לדוגמה בעברית כדי לבדוק את הריווח | he | sample | 67 % | 75 % | sample |
| fi と fl の liga フィーチャーを書いてください | ja | fea | 68 % | 80 % | fea |
| Come faccio a esportare il font in OTF? | it | docs | 69 % | 73 % | docs |
| Ge mig ett pangram på svenska för att testa typsnittet | sv | sample | 71 % | 72 % | sample |
| Napisz coś do okna Glyph, żeby sprawdzić ligatury fi fl | pl | sample | 71 % | 67 % | sample |
| I want a pangram that uses all Cyrillic letters to test in the Glyph window | en | sample | 73 % | 73 % | sample |
| چطور فونت را به صورت OTF خروجی بگیرم؟ | fa | docs | 73 % | 65 % | docs |
| Anna esimerkkiteksti, jossa on paljon ä- ja ö-kirjaimia | fi | sample | 74 % | 72 % | sample |
| Describe the letter O as JSON with nodes for the FontLab format | en | vfj | 75 % | 75 % | vfj |
| Як змінити сімейну назву шрифту? | uk | docs | 76 % | 71 % | docs |
| Escreva o código da feature smcp para versaletes | pt | fea | 76 % | 79 % | fea |
| Jak vyexportovat font jako OTF? | cs | docs | 76 % | 66 % | docs |
| Hoe voeg ik een nieuwe as toe aan een variabel font? | nl | docs | 76 % | 70 % | docs |
| Écris une fonctionnalité liga pour fi et fl | fr | fea | 77 % | 78 % | fea |
| Give me the code that makes ff become a ligature | en | fea | 77 % | 77 % | fea |
| Գրիր kern feature Ա և Վ տառերի համար | hy | fea | 78 % | 79 % | fea |
| Something to type in the preview so I can see all my ligatures | en | sample | 78 % | 78 % | sample |
| The font exports fine but the OTF name shows wrong in Word, why? | en | docs | 79 % | 79 % | docs |
| What would the JSON for a glyph look like? | en | vfj | 79 % | 79 % | vfj |
| Which menu has the option to flip the selection? | en | docs | 80 % | 80 % | docs |
| Írj egy kern feature-t az A és V betűk közé | hu | fea | 80 % | 80 % | fea |
| Write a liga feature that substitutes f i with fi and f l with fl | en | fea | 80 % | 80 % | fea |
| Schreib mir ein kern-Feature für die Buchstaben A V W T | de | fea | 80 % | 74 % | fea |
| Jak zmienić kąt pochylenia kursywy w FontLab? | pl | docs | 81 % | 72 % | docs |
| اكتب كود feature للربط بين الحرفين f و i | ar | fea | 81 % | 84 % | fea |
| Πώς αλλάζω τη γωνία των πλάγιων γραμμάτων στο FontLab; | el | docs | 81 % | 74 % | docs |
| FontLabでイタリック角度を変更するには? | ja | docs | 82 % | 76 % | docs |
| How do I change the italic angle of my font in FontLab? | en | docs | 82 % | 82 % | docs |
| I need kern feature code with classes for round letters versus flat letters | en | fea | 82 % | 82 % | fea |
| Something to type in the Glyph window that shows off my ligatures | en | sample | 82 % | 82 % | sample |
| Wo stelle ich die vertikalen Metriken ein? | de | docs | 82 % | 78 % | docs |
| დამიწერე Python სკრიპტი, რომელიც ყველა გლიფს გაასწორებს | ka | python | 83 % | 92 % | python |
| FontLab में kerning पैनल कहाँ है? | hi | docs | 84 % | 83 % | docs |
| A와 V 사이의 커닝 feature 코드를 작성해 줘 | ko | fea | 84 % | 81 % | fea |
| What does the Match Masters command actually do? | en | docs | 84 % | 84 % | docs |
| 给我一段包含所有拼音字母的示例文本,用来检查字体 | zh | sample | 85 % | 86 % | sample |
| Viết định nghĩa VFJ cho glyph dấu gạch ngang rộng 300 | vi | vfj | 86 % | 90 % | vfj |
| Give me a script to export every open font as OTF into ~/Desktop/out | en | python | 87 % | 87 % | python |
| Дай мне текст-образец с кириллицей для окна Glyph | ru | sample | 89 % | 85 % | sample |
| Dame un texto de muestra con muchas letras redondas | es | sample | 89 % | 90 % | sample |
| Give me a sample text with lots of round letters so I can check my o and c | en | sample | 90 % | 90 % | sample |
| 写一个 Python 脚本,把所有字形的宽度设为 600 | zh | python | 90 % | 92 % | python |
| Escribe un script en Python que elimine los glifos sin unicode | es | python | 90 % | 94 % | python |
| Give me a JSON glyph entry in VFJ format for a composite eacute made of e and acutecomb | en | vfj | 91 % | 91 % | vfj |
| Напиши скрипт на Python, который удалит все направляющие | ru | python | 91 % | 89 % | python |
| Napisz skrypt Python, który zmieni nazwy wszystkich glifów z .sc na .smcp | pl | python | 92 % | 87 % | python |
| Напиши VFJ-описание глифа «точка» с юникодом 002E | bg | vfj | 92 % | 91 % | vfj |
| เขียนสคริปต์ Python ที่เปลี่ยนชื่อ glyph ทั้งหมดเป็นตัวพิมพ์เล็ก | th | python | 93 % | 88 % | python |
| Tüm glifleri 50 birim sağa kaydıran bir Python betiği yaz | tr | python | 93 % | 93 % | python |
| Write a Python script that renames all glyphs ending in .sc to .smcp | en | python | 93 % | 93 % | python |
| একটি VFJ গ্লিফ সংজ্ঞা লিখুন যার নাম period এবং প্রস্থ 250 | bn | vfj | 93 % | 88 % | vfj |
| Write the VFJ definition of a glyph named A with a triangular outline, unicode 0041, width 600 | en | vfj | 95 % | 95 % | vfj |
| Scrivimi la definizione VFJ di un glifo quadrato largo 500 | it | vfj | 95 % | 94 % | vfj |
The translator¶
The translated mode runs Hy-MT2-1.8B at Q4_K_M on llama-server in front of the classifier, with a short glossary of font-editing terms. The table lists the 39 non-English queries, their English translations and the time each took. Each query was translated once, so every classifier saw the same English. Chapter 5 describes the prompt.
| Query | Lang | English | ms |
|---|---|---|---|
| Jak zmienić kąt pochylenia kursywy w FontLab? | pl | How can I change the italic angle in FontLab? | 149 |
| Napisz skrypt Python, który zmieni nazwy wszystkich glifów z .sc na .smcp | pl | Write a Python script that changes the names of all glyphs from .sc to .smcp | 112 |
| Schreib mir ein kern-Feature für die Buchstaben A V W T | de | Write me a kern feature for the letters A V W T | 90 |
| Wo stelle ich die vertikalen Metriken ein? | de | Where do I set the vertical metrics? | 61 |
| Come faccio a esportare il font in OTF? | it | How can I export the font to OTF? | 77 |
| Scrivimi la definizione VFJ di un glifo quadrato largo 500 | it | Write me the VFJ definition for a square glyph with a width of 500 | 99 |
| Écris une fonctionnalité liga pour fi et fl | fr | Write a liga functionality for fi and fl | 70 |
| Donne-moi un pangramme français pour tester mes accents | fr | Give me a French pangram to test my accents | 71 |
| Dame un texto de muestra con muchas letras redondas | es | Give me a sample text with many round letters | 74 |
| Escribe un script en Python que elimine los glifos sin unicode | es | Write a Python script that removes glifos without Unicode. | 85 |
| Jak vyexportovat font jako OTF? | cs | How to export a font as OTF? | 66 |
| Escreva o código da feature smcp para versaletes | pt | Write the code for the smcp feature for Versaletes | 85 |
| Hoe voeg ik een nieuwe as toe aan een variabel font? | nl | How do I add a new font style to a variable font? | 91 |
| Tüm glifleri 50 birim sağa kaydıran bir Python betiği yaz | tr | A Python script that shifts all glyphs 50 units to the right | 95 |
| Írj egy kern feature-t az A és V betűk közé | hu | Write a kern feature between the letters A and V | 81 |
| Anna esimerkkiteksti, jossa on paljon ä- ja ö-kirjaimia | fi | An example text with many A and O letters | 72 |
| Viết định nghĩa VFJ cho glyph dấu gạch ngang rộng 300 | vi | Write the VFJ definition for the wide 300-width horizontal glyph | 90 |
| Ge mig ett pangram på svenska för att testa typsnittet | sv | Give me a pangram in Swedish to test the font | 81 |
| Дай мне текст-образец с кириллицей для окна Glyph | ru | Give me a text sample with Cyrillic characters for the Glyph window | 98 |
| Напиши скрипт на Python, который удалит все направляющие | ru | Write a Python script that removes all leading spaces | 66 |
| Як змінити сімейну назву шрифту? | uk | How do I change the family name of a font? | 83 |
| Напиши VFJ-описание глифа «точка» с юникодом 002E | bg | Write the VFJ description for the glyph “dot” with Unicode 002E | 104 |
| Πώς αλλάζω τη γωνία των πλάγιων γραμμάτων στο FontLab; | el | How do I change the angle of the horizontal lines in FontLab? | 95 |
| Գրիր kern feature Ա և Վ տառերի համար | hy | Griir kern feature A and V for spacing | 81 |
| დამიწერე Python სკრიპტი, რომელიც ყველა გლიფს გაასწორებს | ka | Write the Python script, which contains many glyphs | 76 |
| اكتب كود feature للربط بين الحرفين f و i | ar | Write a feature code to connect the letters f and i | 81 |
| چطور فونت را به صورت OTF خروجی بگیرم؟ | fa | How can I export the font as an OTF file? | 85 |
| תן לי טקסט לדוגמה בעברית כדי לבדוק את הריווח | he | Give me a Hebrew text example so I can check the kerning | 89 |
| پائتھن اسکرپٹ لکھو جو تمام گلفس کی چوڑائی 600 کر دے | ur | Write the font script so that all glyphs have a width of 600. | 104 |
| FontLab में kerning पैनल कहाँ है? | hi | Where is the kerning panel in FontLab? | 77 |
| একটি VFJ গ্লিফ সংজ্ঞা লিখুন যার নাম period এবং প্রস্থ 250 | bn | Write a VFJ glyph definition named period with a width of 250. | 99 |
| เขียนสคริปต์ Python ที่เปลี่ยนชื่อ glyph ทั้งหมดเป็นตัวพิมพ์เล็ก | th | Write a Python script that changes the name of all glyphs to lowercase. | 99 |
| FontLabでイタリック角度を変更するには? | ja | How to change the italic angle in FontLab? | 77 |
| fi と fl の liga フィーチャーを書いてください | ja | Please write the features of fi and fl liga. | 77 |
| 给我一段包含所有拼音字母的示例文本,用来检查字体 | zh | Give me an example text containing all phonetic letters to check the font. | 88 |
| 写一个 Python 脚本,把所有字形的宽度设为 600 | zh | Write a Python script to set the width of all glyphs to 600. | 97 |
| A와 V 사이의 커닝 feature 코드를 작성해 줘 | ko | Write the kerning feature code between A and V | 81 |
| VFJ 형식으로 이름이 hyphen이고 너비가 300인 글리프 정의를 작성해 줘 | ko | The FontLab file has a hyphen name and a glyph definition with a width of 300. | 113 |
| Napisz coś do okna Glyph, żeby sprawdzić ligatury fi fl | pl | Write something in the Glyph window to check the ligatures fi fl | 94 |
Language detection¶
Before it translates, the pipeline has to know whether a query is English. Two detectors, and a majority vote that breaks ties in lingua's favour, were measured over 100 passes of the 67 queries, single-threaded. Chapter 5 lists lingua's four misses.
| Detector | Correct | Mean µs | Calls/s |
|---|---|---|---|
| lingua | 63/67 | 481.1 | 2,079 |
| papagan | 43/67 | 11.1 | 90,437 |
| majority | 63/67 | 492.2 | 2,032 |
Calibration¶
These tables group the answers of decider-0.8b (DreamBlooms' Q8_0 file, the one ornotto registers) on dohnuts (Metal) by their top probability. Share right is the fraction of answers in that band that were correct; a calibrated model is right about as often as its probability says. Chapter 9 reads the bands.
With translation:
| Top probability | Answers | Right | Share right |
|---|---|---|---|
| below 0.6 | 8 | 7 | 88 % |
| 0.6 to 0.8 | 11 | 11 | 100 % |
| 0.8 to 0.95 | 17 | 14 | 82 % |
| 0.95 and up | 31 | 30 | 97 % |
On the original text:
| Top probability | Answers | Right | Share right |
|---|---|---|---|
| below 0.6 | 10 | 8 | 80 % |
| 0.6 to 0.8 | 9 | 8 | 89 % |
| 0.8 to 0.95 | 16 | 14 | 88 % |
| 0.95 and up | 32 | 31 | 97 % |
A fallback gate¶
A gate asks decider-0.8b on dohnuts first and sends the query to Qwen3.5-4B-Hmm on pcdServer only when decider's top probability is below the threshold. The rows replay the benchmark's cached answers at each threshold. Fallbacks counts the queries sent on; mean ms is the cost per query averaged over all 67, where a query that falls back pays for both calls. The thresholds were chosen on the same queries they are scored on, so this is not a held-out test. Chapter 9 explains why the gate pays only on the original text.
With translation:
| Gate: top probability below | Fallbacks | Correct | Mean ms |
|---|---|---|---|
| 0.50 | 7/67 | 62/67 | 61.2 |
| 0.60 | 8/67 | 62/67 | 62.5 |
| 0.70 | 17/67 | 62/67 | 73.9 |
| 0.80 | 19/67 | 62/67 | 76.4 |
| 0.85 | 23/67 | 62/67 | 81.6 |
| 0.90 | 29/67 | 63/67 | 89.2 |
| 0.95 | 36/67 | 63/67 | 98.1 |
On the original text:
| Gate: top probability below | Fallbacks | Correct | Mean ms |
|---|---|---|---|
| 0.50 | 8/67 | 62/67 | 62.7 |
| 0.60 | 10/67 | 62/67 | 65.2 |
| 0.70 | 15/67 | 63/67 | 71.5 |
| 0.80 | 19/67 | 63/67 | 76.6 |
| 0.85 | 22/67 | 63/67 | 80.6 |
| 0.90 | 28/67 | 63/67 | 88.3 |
| 0.95 | 35/67 | 64/67 | 97.2 |