Benchmarks · coding
FrontierCode 2.0
Mergeability of coding-agent changes on private maintainer-authored tasks, graded by unit tests, rubrics and verifiers (Cognition).
28 models · 104 results · data as of 2026-11-06 · cognition.com ↗
Leaderboard
Opus 5.5 max
Heavy line: best score at each price. Thin lines join one model's effort levels. Cost uses each model's median subscription plan, or API price with no plan data. 13 more models under All models in settings.
More from this benchmark
1 published measuresOutput tokens per task104 results · lower is better
13 more models under All models in settings.
All results
104 rows| # | Model | Effort | FrontierCode 2.0 | Subscription $ / task | API $ / task | Output tokens | Harness |
|---|---|---|---|---|---|---|---|
| 1 | Opus 5.5 | max | 45.5% | $0.19 | $13 | 202k | claude-code |
| 2 | Opus 5.5 | xhigh | 43.4% | $0.078 | $5.18 | 88k | claude-code |
| 3 | GPT-6 Astra | max | 42.8% | $0.29 | $6.68 | 36k | codex |
| 4 | GPT-6 Astra | xhigh | 41.4% | $0.19 | $4.54 | 21k | codex |
| 5 | Sonnet 5.5 | max | 41.0% | $0.30 | $20 | 515k | claude-code |
| 6 | Gemini 4 Carbon | medium | 40.0% | $1.44 | $4.04 | 109k | devin-cli |
| 7 | GPT-6.1 Sol | max | 39.0% | $0.060 | $1.03 | 41k | codex |
| 8 | GPT-6 Astra | high | 38.3% | $0.18 | $4.28 | 18k | codex |
| 9 | Gemini 4 Carbon | high | 37.5% | $1.58 | $4.43 | 121k | devin-cli |
| 10 | GPT-6 Astra | medium | 36.8% | $0.15 | $3.49 | 14k | codex |
| 11 | GPT-6.1 Sol | high | 34.8% | $0.038 | $0.65 | 22k | codex |
| 12 | Sonnet 5.5 | xhigh | 34.8% | $0.050 | $3.31 | 80k | claude-code |
| 13 | Opus 5.5 | high | 34.5% | $0.037 | $2.42 | 41k | claude-code |
| 14 | GPT-6 Sol | max | 34.3% | $0.073 | $2.35 | 54k | codex |
| 15 | Fable 5.1 | xhigh | 32.4% | $1.35 | $8.68 | 93k | claude-code |
| 16 | GPT-6.1 Sol | xhigh | 32.3% | $0.041 | $0.69 | 25k | codex |
| 17 | Gemini 4 Carbon | low | 31.9% | $1.62 | $4.55 | 58k | devin-cli |
| 18 | Fable 5.1 | max | 31.7% | $1.79 | $11 | 119k | claude-code |
| 19 | GPT-6 Astra | low | 31.1% | $0.10 | $2.44 | 9.0k | codex |
| 20 | GPT-6 Sol | xhigh | 30.3% | $0.045 | $1.44 | 29k | codex |
| 21 | Opus 5.5 | medium | 30.1% | $0.026 | $1.70 | 30k | claude-code |
| 22 | GPT-6.1 Sol | medium | 29.6% | $0.029 | $0.49 | 16k | codex |
| 23 | Grok 4.7 | xhigh | 28.6% | $1.33 | $14 | 111k | grok-build |
| 24 | Haiku 5.5 | max | 27.7% | $0.008 | $0.37 | 205k | claude-code |
| 25 | GPT-5.6 Sol | max | 27.6% | $0.15 | $5.50 | 38k | codex |
| 26 | Opus 5 | xhigh | 26.7% | $0.19 | $11 | 76k | claude-code |
| 27 | Opus 5 | high | 26.7% | $0.14 | $8.36 | 59k | claude-code |
| 28 | Opus 5 | max | 26.6% | $0.23 | $13 | 88k | claude-code |
| 29 | Grok 4.7 | high | 26.6% | $1.03 | $11 | 94k | grok-build |
| 30 | Gemini 4 Argon | high | 26.3% | $4.64 | $6.51 | 149k | devin-cli |