All models
19 models, sorted by how many use cases each has been run on. Counts include published runs only (not under review); a run that used several models counts for each. Data as of Nov 4, 2026, 20:00 UTC.
| Model | Use cases | Runs | Pass rate | Last run |
|---|---|---|---|---|
| Claude Sonnet 5.5 | 404 | 413 | 87% (361/413) | 2026-10-26 |
| Claude Haiku 5.5 | 404 | 411 | 89% (364/411) | 2026-10-26 |
| Claude Opus 5.5 | 403 | 412 | 92% (379/412) | 2026-10-26 |
| GPT-6.1 Sol | 139 | 139 | 96% (133/139) | 2026-10-23 |
| DeepSeek V4.1 Flash (Venice) | 78 | 82 | 88% (72/82) | 2026-11-04 |
| Qwen3 Coder 480B Turbo (Venice) | 51 | 51 | 55% (28/51) | 2026-11-04 |
| Mistral Small 3.2 24B Instruct (Venice) | 44 | 44 | 7% (3/44) | 2026-10-30 |
| Gemma 4 31B Instruct (Venice) | 42 | 42 | 86% (36/42) | 2026-11-04 |
| Gemini 4 Argon | 40 | 40 | 98% (39/40) | 2026-10-28 |
| Gemini 3.8 Flash | 35 | 35 | 97% (34/35) | 2026-10-23 |
| Grok 4.3 (Venice) | 25 | 25 | 88% (22/25) | 2026-10-19 |
| Llama 3.2 3B (local) | 16 | 16 | 0% (0/16) | 2026-10-20 |
| GLM 4.7 (Venice) | 15 | 15 | 80% (12/15) | 2026-11-04 |
| Qwen2.5 7B (local) | 15 | 15 | 7% (1/15) | 2026-10-20 |
| GLM 5.3 (Venice) | 9 | 9 | 89% (8/9) | 2026-10-17 |
| Kimi K3.1 Pro (Venice) | 9 | 9 | 89% (8/9) | 2026-10-30 |
| MiniMax M2.5 (Venice) | 5 | 5 | 80% (4/5) | 2026-10-27 |
| DeepSeek V3.2 (Venice) | 4 | 4 | 25% (1/4) | 2026-10-28 |
| Kimi K2.5 (Venice) | 4 | 4 | 50% (2/4) | 2026-10-24 |