FrontierCode 1.1 Main
We show this table for reference; we do not rank on it.
Cognition's 100-task software-engineering benchmark for whether coding agents produce mergeable, production-quality pull requests, scored for correctness, tests, scope, style, and maintainability through maintainer-authored rubrics.
Main score on FrontierCode 1.1 Main — October 23, 2026 snapshot
We mirror the published main score view for FrontierCode 1.1 Main. Claude Opus 5.5 leads the public snapshot at 54.6%, followed by Claude Fable 5 (53.5%) and Claude Opus 5 (53.4%). We do not use these results to rank models overall.
Claude Opus 5.5
Anthropic
claude-code · medium reasoning
Claude Fable 5
Anthropic
claude-code · xhigh reasoning
Claude Opus 5
Anthropic
claude-code · medium reasoning
46 modelsCoding4% of Coding reference weightCurrentUpdated October 23, 2026 snapshot
Main score table (46 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Claude Opus 5.5Anthropicclaude-code · medium reasoning | 54.6% | Not reported | Closed |
| 2 | Claude Fable 5Anthropicclaude-code · xhigh reasoning | 53.5% | Not reported | Closed |
| 3 | Claude Opus 5Anthropicclaude-code · medium reasoning | 53.4% | Not reported | Closed |
| 4 | GPT-6 AstraOpenAIcodex · max reasoning | 53.3% | Not reported | Closed |
| 5 | Claude Sonnet 5.5Anthropicclaude-code · xhigh reasoning | 52.1% | Not reported | Closed |
| 6 | Claude Fable 5.1Anthropicclaude-code · medium reasoning | 50.9% | Not reported | Closed |
| 7 | GPT-6.1 SolOpenAIcodex · medium reasoning | 50.2% | Not reported | Closed |
| 8 | SWE-2Cognitiondevin · max reasoning | 50.0% | Not reported | Closed |
| 9 | GPT-6 SolOpenAIcodex · max reasoning | 49.3% | Not reported | Closed |
| 10 | Grok 4.6xAIgrok-build · high reasoning | 48.0% | Not reported | Closed |
| 11 | Grok 4.7xAIgrok-build · high reasoning | 47.6% | Not reported | Closed |
| 12 | Gemini 4 ArgonGooglechisel · medium reasoning | 47.5% | Not reported | Closed |
| 13 | GPT-5.6 SolOpenAIcodex · max reasoning | 47.5% | Not reported | Closed |
| 14 | Claude Opus 4.8Anthropicclaude-code · max reasoning | 46.5% | Not reported | Closed |
| 15 | Claude Haiku 5.5Anthropicclaude-code · max reasoning | 46.4% | Not reported | Closed |
| 16 | Kimi K3Moonshot AImini-swe-agent · none reasoning | 44.2% | Not reported | Pending |
| 17 | DeepSeek V4.1 Flash 1017DeepSeekchisel · high reasoning | 43.9% | Not reported | Open |
| 18 | Gemini 3.7 FlashGooglechisel · medium reasoning | 43.6% | Not reported | Closed |
| 19 | GPT-5.5OpenAIcodex · xhigh reasoning | 43.0% | Not reported | Closed |
| 20 | Claude Sonnet 5Anthropicclaude-code · xhigh reasoning | 42.7% | Not reported | Closed |
| 21 | GPT-6 LunaOpenAIcodex · max reasoning | 42.4% | Not reported | Closed |
| 22 | Grok 4.5xAIgrok-build · high reasoning | 42.4% | Not reported | Closed |
| 23 | SWE-1.7Cognitionchisel · none reasoning | 42.0% | Not reported | Closed |
| 24 | GPT-5.6 TerraOpenAIcodex · max reasoning | 41.3% | Not reported | Closed |
| 25 | Gemini 3.8 FlashGooglechisel · medium reasoning | 41.2% | Not reported | Closed |
| 26 | GLM-5.3Z.AIchisel · max reasoning | 40.1% | Not reported | Open |
| 27 | GPT-5.6 LunaOpenAIcodex · max reasoning | 39.8% | Not reported | Closed |
| 28 | Mistral Large 4Mistralchisel · high reasoning | 39.4% | Not reported | Pending |
| 29 | Claude Opus 4.7Anthropicclaude-code · max reasoning | 38.5% | Not reported | Closed |
| 30 | Gemini 3.6 FlashGooglechisel · medium reasoning | 34.4% | Not reported | Closed |
| 31 | GLM-5.3-FlashZ.AIchisel · max reasoning | 31.8% | Not reported | Open |
| 32 | Kimi K2.7Moonshot AImini-swe-agent · none reasoning | 30.1% | Not reported | Unknown |
| 33 | DeepSeek V4 Pro 0813DeepSeekchisel · high reasoning | 28.5% | Not reported | Open |
| 34 | GPT-5.4 miniOpenAIcodex · xhigh reasoning | 27.0% | Not reported | Closed |
| 35 | Claude Opus 4.6Anthropicclaude-code · high reasoning | 26.6% | Not reported | Closed |
| 36 | Composer 2.5Cursorcursor-cli · none reasoning | 25.6% | Not reported | Closed |
| 37 | GLM-5.2Z.AImini-swe-agent · none reasoning | 24.5% | Not reported | Open |
| 38 | Claude Sonnet 4.6Anthropicclaude-code · max reasoning | 24.3% | Not reported | Closed |
| 39 | DeepSeek V4 Flash 0731DeepSeekchisel · high reasoning | 18.8% | Not reported | Open |
| 40 | DeepSeek V4 ProDeepSeekmini-swe-agent · none reasoning | 17.6% | Not reported | Unknown |
| 41 | MiniMax M3MiniMaxmini-swe-agent · none reasoning | 14.7% | Not reported | Open |
| 42 | InklingThinking Machines Labmini-swe-agent · 0.99 reasoning | 14.0% | Not reported | Open |
| 43 | Nemotron 3 UltraNVIDIAchisel · none reasoning | 13.6% | Not reported | Open |
| 44 | Qwen3.7 PlusAlibabamini-swe-agent · none reasoning | 10.2% | Not reported | Closed |
| 45 | SWE-1.6Cognitionchisel · none reasoning | 9.4% | Not reported | Unknown |
| 46 | Mistral 3.5 MediumMistralchisel · none reasoning | 8.0% | Not reported | Unknown |
How we show FrontierCode 1.1 Main
The snapshot mirrors Cognition's FrontierCode 1.1 Main table captured on October 23, 2026 snapshot. Main uses the 100 hardest tasks from the 150-task Extended set. Each table selects the best published reasoning effort for its own task set.
Each row keeps the harness and settings the source published. BenchLM shows the published table for reference; the per-model scores it stores feed the ranking.
Snapshot
The published FrontierCode 1.1 Main snapshot places Claude Opus 5.5 first at 54.6%. The third row is 1.2 points behind. The broader top-10 range is 6.6 points, so many of the published results sit in a relatively narrow band.
46 models have been evaluated on FrontierCode 1.1 Main. The benchmark falls in the Coding category. BenchLM shows the published table for reference; the per-model scores it stores feed the ranking. BenchAlign v5.8 gives FrontierCode 1.1 Main 4% of the Coding reference weight, so it moves the Coding leaderboard and the overall ranking. Reference weights are relative weights in the calibrated model, not fixed shares of a score.
About FrontierCode 1.1 Main
Year
2026
Tasks
100 private Main tasks (150 in Extended)
Format
Repository task completion with maintainer rubrics
Difficulty
Frontier coding-agent quality
FrontierCode 1.1 Main uses 100 of the benchmark's 150 private software-engineering tasks. The leaderboard reports the best-performing published reasoning effort for each model-agent row. Each row combines a model with an agent harness, and the private tasks cannot be independently rerun from the public artifact.
Freshness and provenance
Version
FrontierCode 1.1 Main
Refresh cadence
Rolling
Staleness state
Current
Question availability
Private tasks with public aggregate results
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does FrontierCode 1.1 Main measure?
Cognition's 100-task software-engineering benchmark for whether coding agents produce mergeable, production-quality pull requests, scored for correctness, tests, scope, style, and maintainability through maintainer-authored rubrics.
Which model leads the published FrontierCode 1.1 Main snapshot?
Claude Opus 5.5 currently leads the published FrontierCode 1.1 Main snapshot with 54.6% main score. BenchLM shows the published table for reference; the per-model scores it stores feed the ranking.
How many models are evaluated on FrontierCode 1.1 Main?
The October 23, 2026 snapshot snapshot contains 46 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 5,600+ readers.
One email each week. Unsubscribe anytime.