| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | GPT-5.6 Sol | OpenAI | 89.4 | 1 | 2026-07-16 |
| 2 | Claude Fable 5 | Anthropic | 88 | 1 | 2026-07-16 |
| 3 | Gemini 3.1 Pro | 85.9 | 1 | 2026-07-16 | |
| 4 | DeepSeek-V4-Pro | DeepSeek | 83.4 | 1 | 2026-07-16 |
| 5 | Kimi K2.6 | Moonshot | 83.2 | 1 | 2026-07-16 |
| 6 | Inkling | Thinking Machines | 77.1 | 1 | 2026-07-16 |
| 7 | Kimi K2.5 | Moonshot | 74.9 | 1 | 2026-07-16 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.