Benchmarks · coding
FrontierCode 1.1 Main
Mergeability of coding-agent changes on private maintainer-authored tasks, graded by unit tests, rubrics and verifiers (Cognition).
45 models · 134 results · data as of 2026-10-29 · cognition.com ↗
Leaderboard
Opus 5.5 medium
More from this benchmark
2 published measuresTime per task24 results · lower is better
Output tokens per task134 results · lower is better
All results
134 rows| # | Model | Effort | FrontierCode 1.1 Main | Subscription $ / task | API $ / task | Time | Output tokens | Harness |
|---|---|---|---|---|---|---|---|---|
| 1 | Opus 5.5 | medium | 54.6% | $0.012 | $0.80 | 6.3 min | 19k | claude-code |
| 2 | Opus 5.5 | max | 54.4% | $0.094 | $6.19 | 25 min | 166k | claude-code |
| 3 | Opus 5.5 | high | 54.0% | $0.016 | $1.09 | 7.2 min | 26k | claude-code |
| 4 | Fable 5 | xhigh | 53.5% | $1.48 | $13 | 59k | claude-code | |
| 5 | Opus 5 | medium | 53.4% | $0.075 | $4.31 | 34k | claude-code | |
| 6 | GPT-6 Astra | max | 53.3% | $0.39 | $4.59 | 30k | codex | |
| 7 | Fable 5 | high | 52.7% | $1.07 | $9.48 | 43k | claude-code | |
| 8 | Sonnet 5.5 | xhigh | 52.1% | $0.024 | $1.59 | 13 min | 47k | claude-code |
| 9 | Fable 5 | max | 51.6% | $2.15 | $19 | 81k | claude-code | |
| 10 | Opus 5.5 | xhigh | 51.4% | $0.034 | $2.25 | 11 min | 60k | claude-code |
| 11 | GPT-6 Astra | high | 50.9% | $0.26 | $3.01 | 14k | codex | |
| 12 | Fable 5.1 | medium | 50.9% | $0.51 | $3.28 | 26k | claude-code | |
| 13 | GPT-6 Astra | xhigh | 50.6% | $0.28 | $3.28 | 17k | codex | |
| 14 | Fable 5.1 | high | 50.3% | $0.82 | $5.27 | 40k | claude-code | |
| 15 | Fable 5.1 | max | 50.3% | $2.00 | $13 | 92k | claude-code | |
| 16 | GPT-6.1 Sol | medium | 50.2% | $0.042 | $0.36 | 14 min | 11k | codex |
| 17 | SWE 2 | max | 50.0% | $1.18 | $1.18 | 73k | devin | |
| 18 | Fable 5.1 | low | 49.8% | $0.37 | $2.38 | 19k | claude-code | |
| 19 | Fable 5 | medium | 49.8% | $0.81 | $7.15 | 33k | claude-code | |
| 20 | Sonnet 5.5 | high | 49.4% | $0.006 | $0.42 | 5.9 min | 13k | claude-code |
| 21 | GPT-6.1 Sol | xhigh | 49.3% | $0.065 | $0.56 | 17 min | 19k | codex |
| 22 | GPT-6 Sol | max | 49.3% | $0.13 | $2.07 | 41k | codex | |
| 23 | GPT-6 Astra | medium | 48.8% | $0.21 | $2.43 | 11k | codex | |
| 24 | Fable 5.1 | xhigh | 48.7% | $1.45 | $9.27 | 68k | claude-code | |
| 25 | GPT-6 Sol | xhigh | 48.4% | $0.082 | $1.32 | 23k | codex | |
| 26 | Opus 5 | max | 48.0% | $0.20 | $11 | 98k | claude-code | |
| 27 | Fable 5 | low | 48.0% | $0.55 | $4.92 | 24k | claude-code | |
| 28 | Grok 4.6 | high | 48.0% | $0.30 | $2.88 | 37k | grok-build | |
| 29 | Opus 5 | high | 48.0% | $0.13 | $7.24 | 55k | claude-code | |
| 30 | GPT-6.1 Sol | high | 48.0% | $0.058 | $0.50 | 16 min | 16k | codex |