Benchmarks · coding
FrontierCode 2.0
Uses the 2.0 Main suite's weighted rubric score at each model's best reasoning effort, including published historical model results. Score: Main rubric score at best reasoning effort. Collected by Jake Boggs' benchmark dashboard.
27 models · 27 results · data as of 2026-11-06 · cognition.com ↗
Leaderboard
Opus 5.5 max
All results
27 rows| # | Model | Effort | FrontierCode 2.0 |
|---|---|---|---|
| 1 | Opus 5.5 | max | 45.5% |
| 2 | GPT-6 Astra | max | 42.8% |
| 3 | Sonnet 5.5 | max | 41.0% |
| 4 | Gemini 4 Carbon | max | 40.0% |
| 5 | GPT-6.1 Sol | max | 39.0% |
| 6 | GPT-6 Sol | max | 34.3% |
| 7 | Fable 5.1 | max | 32.4% |
| 8 | Grok 4.7 | max | 28.6% |
| 9 | Haiku 5.5 | max | 27.7% |
| 10 | GPT-5.6 Sol | max | 27.6% |
| 11 | Opus 5 | max | 26.7% |
| 12 | Gemini 4 Argon | max | 26.3% |
| 13 | Muse Wonder (Max) | max | 24.3% |
| 14 | SWE 2 Flash | max | 24.0% |
| 15 | SWE 2 | max | 23.8% |
| 16 | GPT-5.5 | xhigh | 22.6% |
| 17 | GPT-6 Luna | max | 21.9% |
| 18 | Mistral Large 4 | max | 19.7% |
| 19 | DeepSeek V4.1 Flash 1017 | max | 19.3% |
| 20 | GLM 5.3 | max | 17.8% |
| 21 | Kimi K3 | max | 17.7% |
| 22 | SWE 1.7 | none | 16.2% |
| 23 | Opus 4.8 | max | 14.7% |
| 24 | GLM 5.3 Flash | max | 14.2% |
| 25 | Muse Spark Open | xhigh | 6.2% |
| 26 | Inkling Small | max | 3.2% |
| 27 | SWE 1.6 | none | 1.3% |