CursorBench 4.0: leaderboard
Cursor coding-agent benchmark for ambiguous, multi-file real-world tasks; the 4.0 task set adds long-horizon editing, refactoring, investigation, intent understanding, job management, and design adherence.
Metric: Score (%). Source: cursor.com. Source: cursor.com. Source: epoch.ai. Saturation forecast: Around 2029. 81 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (Max) | 57.8 | #3 (Claude Opus 5.5) |
| 2 | Claude Opus 5.5 (High) | 56 | #3 (Claude Opus 5.5) |
| 3 | Claude Opus 5.5 (xHigh) | 56 | #3 (Claude Opus 5.5) |
| 4 | Claude Sonnet 5.5 (Max) | 55.5 | #14 (Claude Sonnet 5.5) |
| 5 | Claude Sonnet 5.5 (xHigh) | 53.1 | #14 (Claude Sonnet 5.5) |
| 6 | Gemini 4 Argon (High) | 52.9 | #9 (Gemini 4 Argon) |
| 7 | Gemini 4 Argon (Medium) | 52.8 | #9 (Gemini 4 Argon) |
| 8 | Claude Opus 5.5 (Medium) | 52.5 | #3 (Claude Opus 5.5) |
| 9 | Claude Fable 5.1 (Max) | 51.8 | #10 (Claude Fable 5.1) |
| 10 | Claude Fable 5.1 (xHigh) | 51.6 | #10 (Claude Fable 5.1) |
| 11 | Gemini 4 Argon (Low) | 50.6 | #9 (Gemini 4 Argon) |
| 12 | Muse Wonder (Max) | 50.4 | #18 (Muse Wonder) |
| 13 | Claude Fable 5.1 (High) | 49.2 | #10 (Claude Fable 5.1) |
| 14 | Claude Haiku 5.5 (Max) | 48.4 | #43 (Claude Haiku 5.5) |
| 15 | Muse Wonder (xHigh) | 48.2 | #18 (Muse Wonder) |
No result here: #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=cursorbench-4-0 · How It Works · Data refreshed daily, snapshot 2026-11-05.