CursorBench
ReportedNo description yet: we only describe an eval set when we can do it correctly.
- Kind
- Published by a third party's leaderboard, collected by Epoch AI. The leaderboard is named and linked.
- Link
- https://cursor.com/cursorbench
- Score shown
- Epoch's column "Score", as a percentage, higher is better. The source gives no error bars.
- Results
- 94 in all: 10 for open-weight models, 84 for closed ones. 10 linked to a model page here.
Best open-weight result: deepseek-ai/DeepSeek-V4.1-Flash-1017, 47.7%, number 28 of 94 results.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | gemini-4-carbon_highGoogle DeepMind | closed | effort high; Reasoning level: High | 60.4% | not dated | cursor.com |
| 2 | gemini-4-carbon_mediumGoogle DeepMind | closed | effort medium; Reasoning level: Medium | 60.1% | not dated | cursor.com |
| 3 | claude-fable-5-2_maxAnthropic | closed | effort max; Reasoning level: Max | 60.0% | not dated | cursor.com |
| 4 | gemini-4-carbon_lowGoogle DeepMind | closed | effort low; Reasoning level: Low | 58.6% | not dated | cursor.com |
| 5 | claude-opus-5-5_maxAnthropic | closed | effort max; Reasoning level: Max | 57.8% | not dated | cursor.com |
| 6 | gemini-4-hydrogen_xhighGoogle DeepMind | closed | effort xhigh; Reasoning level: Extra High | 57.8% | not dated | cursor.com |
| 7 | claude-fable-5-2_highAnthropic | closed | effort high; Reasoning level: High | 56.9% | not dated | cursor.com |
| 8 | claude-fable-5-2_xhighAnthropic | closed | effort xhigh; Reasoning level: Extra High | 56.3% | not dated | cursor.com |
| 9 | claude-opus-5-5_highAnthropic | closed | effort high; Reasoning level: High | 56.0% | not dated | cursor.com |
| 10 | claude-opus-5-5_xhighAnthropic | closed | effort xhigh; Reasoning level: Extra High | 56.0% | not dated | cursor.com |
| 11 | claude-sonnet-5-5_maxAnthropic | closed | effort max; Reasoning level: Max | 55.5% | not dated | cursor.com |
| 12 | claude-fable-5-2_mediumAnthropic | closed | effort medium; Reasoning level: Medium | 54.8% | not dated | cursor.com |
| 13 | gemini-4-hydrogen_highGoogle DeepMind | closed | effort high; Reasoning level: High | 54.3% | not dated | cursor.com |
| 14 | claude-fable-5-2_lowAnthropic | closed | effort low; Reasoning level: Low | 53.2% | not dated | cursor.com |
| 15 | claude-sonnet-5-5_xhighAnthropic | closed | effort xhigh; Reasoning level: Extra High | 53.1% | not dated | cursor.com |
| 16 | gemini-4-argon_highGoogle DeepMind | closed | effort high; Reasoning level: High | 52.9% | not dated | cursor.com |
| 17 | gemini-4-argon_mediumGoogle DeepMind | closed | effort medium; Reasoning level: Medium | 52.8% | not dated | cursor.com |
| 18 | claude-opus-5-5_mediumAnthropic | closed | effort medium; Reasoning level: Medium | 52.5% | not dated | cursor.com |
| 19 | claude-fable-5-1_maxAnthropic | closed | effort max; Reasoning level: Max | 51.8% | not dated | cursor.com |
| 20 | claude-fable-5-1_xhighAnthropic | closed | effort xhigh; Reasoning level: Extra High | 51.6% | not dated | cursor.com |
| 21 | gemini-4-argon_lowGoogle DeepMind | closed | effort low; Reasoning level: Low | 50.6% | not dated | cursor.com |
| 22 | muse-wonder-1.0_maxMeta AI | closed | effort max; Reasoning level: Max | 50.4% | not dated | cursor.com |
| 23 | claude-fable-5-1_highAnthropic | closed | effort high; Reasoning level: High | 49.2% | not dated | cursor.com |
| 24 | gemini-4-argon_minimalGoogle DeepMind | closed | effort minimal | 48.8% | not dated | cursor.com |
| 25 | claude-haiku-5-5_maxAnthropic | closed | effort max; Reasoning level: Max | 48.4% | not dated | cursor.com |
| 26 | muse-wonder-1.0_xhighMeta AI | closed | effort xhigh; Reasoning level: Extra High | 48.2% | not dated | cursor.com |
| 27 | claude-sonnet-5-5_highAnthropic | closed | effort high; Reasoning level: High | 47.8% | not dated | cursor.com |
| 28 | deepseek-ai/DeepSeek-V4.1-Flash-1017Epoch: deepseek-v4.1-flash-1017_max, DeepSeek | open | effort max; Reasoning level: Max | 47.7% | not dated | cursor.com |
| 29 | claude-fable-5-1_mediumAnthropic | closed | effort medium; Reasoning level: Medium | 46.8% | not dated | cursor.com |
| 30 | gemini-4-hydrogen_mediumGoogle DeepMind | closed | effort medium; Reasoning level: Medium | 46.7% | not dated | cursor.com |
| 31 | claude-opus-5_maxAnthropic | closed | effort max; Reasoning level: Max | 46.6% | not dated | cursor.com |
| 32 | deepseek-ai/DeepSeek-V4.1-Flash-1017Epoch: deepseek-v4.1-flash-1017_medium, DeepSeek | open | effort medium; Reasoning level: Medium | 46.3% | not dated | cursor.com |
| 33 | grok-4.7_xhighxAI | closed | effort xhigh; Reasoning level: Extra High | 46.3% | not dated | cursor.com |
| 34 | claude-opus-5_xhighAnthropic | closed | effort xhigh; Reasoning level: Extra High | 46.1% | not dated | cursor.com |
| 35 | muse-wonder-1.0_highMeta AI | closed | effort high; Reasoning level: High | 46.1% | not dated | cursor.com |
| 36 | deepseek-ai/DeepSeek-V4.1-Flash-1017Epoch: deepseek-v4.1-flash-1017_high, DeepSeek | open | effort high; Reasoning level: High | 45.6% | not dated | cursor.com |
| 37 | muse-wonder-1.0_mediumMeta AI | closed | effort medium; Reasoning level: Medium | 45.2% | not dated | cursor.com |
| 38 | claude-fable-5-1_lowAnthropic | closed | effort low; Reasoning level: Low | 45.1% | not dated | cursor.com |
| 39 | deepseek-ai/DeepSeek-V4.1-Flash-1017Epoch: deepseek-v4.1-flash-1017_low, DeepSeek | open | effort low; Reasoning level: Low | 44.9% | not dated | cursor.com |
| 40 | claude-opus-5_highAnthropic | closed | effort high; Reasoning level: High | 44.7% | not dated | cursor.com |
| 41 | claude-haiku-5-5_xhighAnthropic | closed | effort xhigh; Reasoning level: Extra High | 44.3% | not dated | cursor.com |
| 42 | grok-4.7_highxAI | closed | effort high; Reasoning level: High | 43.9% | not dated | cursor.com |
| 43 | claude-opus-5-5_lowAnthropic | closed | effort low; Reasoning level: Low | 43.7% | not dated | cursor.com |
| 44 | claude-opus-5_mediumAnthropic | closed | effort medium; Reasoning level: Medium | 43.3% | not dated | cursor.com |
| 45 | zai-org/GLM-5.3Epoch: glm-5.3_max, Z.ai (Zhipu AI) | open | effort max; Reasoning level: Max | 42.6% | not dated | cursor.com |
| 46 | claude-haiku-5-5_highAnthropic | closed | effort high; Reasoning level: High | 42.3% | not dated | cursor.com |
| 47 | gpt-5.6-sol_maxOpenAI | closed | effort max; Reasoning level: Max | 41.7% | not dated | cursor.com |
| 48 | muse-wonder-1.0_lowMeta AI | closed | effort low; Reasoning level: Low | 41.7% | not dated | cursor.com |
| 49 | grok-4.7_mediumxAI | closed | effort medium; Reasoning level: Medium | 41.6% | not dated | cursor.com |
| 50 | muse-spark-1.3_maxMeta AI | closed | effort max; Reasoning level: Max | 41.6% | not dated | cursor.com |
| 51 | grok-4.6_xhighxAI | closed | effort xhigh; Reasoning level: Extra High | 41.4% | not dated | cursor.com |
| 52 | gpt-5.6-terra_maxOpenAI | closed | effort max; Reasoning level: Max | 41.3% | not dated | cursor.com |
| 53 | claude-opus-5_lowAnthropic | closed | effort low; Reasoning level: Low | 40.7% | not dated | cursor.com |
| 54 | grok-4.6_highxAI | closed | effort high; Reasoning level: High | 40.4% | not dated | cursor.com |
| 55 | gemini-3.8-flash_highGoogle DeepMind | closed | effort high; Reasoning level: High | 39.6% | not dated | cursor.com |
| 56 | claude-sonnet-5-5_mediumAnthropic | closed | effort medium; Reasoning level: Medium | 39.2% | not dated | cursor.com |
| 57 | zai-org/GLM-5.3Epoch: glm-5.3_high, Z.ai (Zhipu AI) | open | effort high; Reasoning level: High | 38.0% | not dated | cursor.com |
| 58 | gpt-5.6-sol_xhighOpenAI | closed | effort xhigh; Reasoning level: Extra High | 37.7% | not dated | cursor.com |
| 59 | muse-spark-1.3_xhighMeta AI | closed | effort xhigh; Reasoning level: Extra High | 37.5% | not dated | cursor.com |
| 60 | gemini-3.8-flash_mediumGoogle DeepMind | closed | effort medium; Reasoning level: Medium | 37.3% | not dated | cursor.com |
| 61 | claude-haiku-5-5_mediumAnthropic | closed | effort medium; Reasoning level: Medium | 36.9% | not dated | cursor.com |
| 62 | zai-org/GLM-5.3-FlashEpoch: glm-5.3-flash_max, Z.ai (Zhipu AI) | open | effort max; Reasoning level: Max | 36.8% | not dated | cursor.com |
| 63 | grok-4.6_mediumxAI | closed | effort medium; Reasoning level: Medium | 36.1% | not dated | cursor.com |
| 64 | gpt-5.6-luna_maxOpenAI | closed | effort max; Reasoning level: Max | 35.9% | not dated | cursor.com |
| 65 | claude-sonnet-5-5_lowAnthropic | closed | effort low; Reasoning level: Low | 35.8% | not dated | cursor.com |
| 66 | gpt-5.6-sol_highOpenAI | closed | effort high; Reasoning level: High | 35.7% | not dated | cursor.com |
| 67 | muse-wonder-1.0_minimalMeta AI | closed | effort minimal | 34.2% | not dated | cursor.com |
| 68 | claude-sonnet-5_maxAnthropic | closed | effort max; Reasoning level: Max | 34.1% | not dated | cursor.com |
| 69 | gpt-5.6-terra_xhighOpenAI | closed | effort xhigh; Reasoning level: Extra High | 33.6% | not dated | cursor.com |
| 70 | grok-4.6_lowxAI | closed | effort low; Reasoning level: Low | 33.4% | not dated | cursor.com |
| 71 | muse-spark-1.3_highMeta AI | closed | effort high; Reasoning level: High | 33.4% | not dated | cursor.com |
| 72 | zai-org/GLM-5.3Epoch: glm-5.3_low, Z.ai (Zhipu AI) | open | effort low; Reasoning level: Low | 33.3% | not dated | cursor.com |
| 73 | grok-4.7_lowxAI | closed | effort low; Reasoning level: Low | 33.1% | not dated | cursor.com |
| 74 | gpt-5.6-luna_xhighOpenAI | closed | effort xhigh; Reasoning level: Extra High | 33.0% | not dated | cursor.com |
| 75 | muse-spark-1.3_mediumMeta AI | closed | effort medium; Reasoning level: Medium | 32.6% | not dated | cursor.com |
| 76 | claude-sonnet-5_xhighAnthropic | closed | effort xhigh; Reasoning level: Extra High | 32.0% | not dated | cursor.com |
| 77 | zai-org/GLM-5.3-FlashEpoch: glm-5.3-flash_high, Z.ai (Zhipu AI) | open | effort high; Reasoning level: High | 31.1% | not dated | cursor.com |
| 78 | gpt-5.6-sol_mediumOpenAI | closed | effort medium; Reasoning level: Medium | 31.1% | not dated | cursor.com |
| 79 | claude-haiku-5-5_lowAnthropic | closed | effort low; Reasoning level: Low | 30.9% | not dated | cursor.com |
| 80 | claude-sonnet-5_highAnthropic | closed | effort high; Reasoning level: High | 30.8% | not dated | cursor.com |
| 81 | gpt-5.6-terra_highOpenAI | closed | effort high; Reasoning level: High | 30.7% | not dated | cursor.com |
| 82 | gpt-5.6-luna_highOpenAI | closed | effort high; Reasoning level: High | 29.4% | not dated | cursor.com |
| 83 | muse-spark-1.3_lowMeta AI | closed | effort low; Reasoning level: Low | 29.3% | not dated | cursor.com |
| 84 | claude-sonnet-5_mediumAnthropic | closed | effort medium; Reasoning level: Medium | 28.0% | not dated | cursor.com |
| 85 | gemini-4-hydrogen_lowGoogle DeepMind | closed | effort low; Reasoning level: Low | 27.9% | not dated | cursor.com |
| 86 | Composer 2.5Cursor | closed | default | 27.7% | not dated | cursor.com |
| 87 | gpt-5.6-terra_mediumOpenAI | closed | effort medium; Reasoning level: Medium | 27.6% | not dated | cursor.com |
| 88 | zai-org/GLM-5.3-FlashEpoch: glm-5.3-flash_low, Z.ai (Zhipu AI) | open | effort low; Reasoning level: Low | 26.9% | not dated | cursor.com |
| 89 | gpt-5.6-terra_lowOpenAI | closed | effort low; Reasoning level: Low | 25.2% | not dated | cursor.com |
| 90 | gpt-5.6-sol_lowOpenAI | closed | effort low; Reasoning level: Low | 24.6% | not dated | cursor.com |
| 91 | muse-spark-1.3_minimalMeta AI | closed | effort minimal | 24.3% | not dated | cursor.com |
| 92 | claude-sonnet-5_lowAnthropic | closed | effort low; Reasoning level: Low | 24.1% | not dated | cursor.com |
| 93 | gpt-5.6-luna_mediumOpenAI | closed | effort medium; Reasoning level: Medium | 22.2% | not dated | cursor.com |
| 94 | gpt-5.6-luna_lowOpenAI | closed | effort low; Reasoning level: Low | 16.0% | not dated | cursor.com |
Open-weight rows are shaded. Dates are when the result was run or evaluated, where the data says. Data from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Epoch AI. Method