LLM Stats (FrontierCode 1.1): leaderboard
Coding-agent benchmark testing whether agent-produced code changes are mergeable, judged by unit tests, maintainer-defined rubrics, and verifiers; runs flagged for unfair internet use score zero.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around September 2027. 23 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 | 54.4 | #4 |
| 2 | Claude Fable 5 | 53.5 | #16 |
| 3 | Claude Opus 5 | 53.4 | #12 |
| 4 | GPT-6 Astra | 53.3 | #6 |
| 5 | Claude Sonnet 5.5 | 52.1 | #11 |
| 6 | GPT-6.1 Sol | 50.2 | #7 |
| 7 | GPT-6 Sol | 49.3 | #15 |
| 8 | GPT-5.6 Sol | 47.5 | #18 |
| 9 | Claude Opus 4.8 | 46.5 | #35 |
| 10 | Claude Haiku 5.5 | 46.4 | #37 |
| 11 | Gemini 3.7 Flash | 43.6 | #22 |
| 12 | GPT-5.5 | 43 | #29 |
| 13 | Claude Sonnet 5 | 42.7 | #57 |
| 14 | Grok 4.5 | 42.4 | #49 |
| 15 | GPT-6 Luna | 42.4 | #60 |
No result here: #8 Gemini 4 Argon, #9 Claude Fable 5.1, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-frontiercode-1-1 · How It Works · Data refreshed daily, snapshot 2026-10-23.