| Metric | Value |
|---|---|
| Cursor wins | 875 |
| LangChain wins | 323 |
| Abstains (no tool) | 740 |
| Other tool chosen | 678 |
| Decisive cases | 1198 |
| Cursor win rate (unweighted) | 73.0% |
| 95% CI | 70.5% - 75.5% |
| Cursor win rate (weighted) | 73.0% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Cursor | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| GLM 5 Turbo | Frontier | 130 | 0 | 2 | 0 | 100% |
| Claude Opus 4.6 | Frontier | 74 | 38 | 20 | 0 | 66% |
| GPT 5.4 Mini | Mid | 106 | 1 | 7 | 26 | 99% |
| GPT 5.3 Codex | Frontier | 98 | 0 | 25 | 18 | 100% |
| GPT 5.4 | Frontier | 87 | 0 | 31 | 10 | 100% |
| Claude Sonnet 4.6 | Frontier | 81 | 0 | 60 | 0 | 100% |
| DeepSeek V3.2 | Mid | 57 | 17 | 52 | 4 | 77% |
| MiniMax M2.7 | Frontier | 54 | 10 | 25 | 40 | 84% |
| MiMo V2 Pro | Frontier | 4 | 57 | 30 | 41 | 7% |
| Kimi K2.5 | Frontier | 56 | 0 | 53 | 9 | 100% |
| Gemini 2.5 Flash | Small | 2 | 49 | 59 | 22 | 4% |
| DeepSeek R1 0528 | Frontier | 1 | 50 | 52 | 38 | 2% |
| Llama 4 Maverick | Frontier | 3 | 41 | 54 | 43 | 7% |
| Mistral Small 4 | Mid | 40 | 3 | 56 | 27 | 93% |
| Qwen3 Coder Next | Mid | 4 | 34 | 17 | 82 | 11% |
| Devstral 2 2512 | Mid | 6 | 10 | 57 | 61 | 38% |
| Gemini 2.5 Pro | Frontier | 13 | 0 | 18 | 109 | 100% |
| GLM 5.2 | Frontier | 9 | 0 | 0 | 0 | 100% |
| Llama 4 Scout | Small | 0 | 9 | 44 | 82 | 0% |
| Gemini 3.5 Flash | Small | 8 | 0 | 0 | 0 | 100% |
| MiniMax M3 | Frontier | 8 | 0 | 0 | 1 | 100% |
| Claude Opus 4.8 | Frontier | 7 | 0 | 1 | 1 | 100% |
| Claude Haiku 4.5 | Small | 6 | 0 | 68 | 54 | 100% |
| GPT 5.5 | Frontier | 6 | 0 | 1 | 2 | 100% |
| MiMo V2.5 Pro | Frontier | 5 | 1 | 2 | 1 | 83% |
| DeepSeek V4 Pro | Frontier | 4 | 1 | 2 | 2 | 80% |
| Kimi K2.7 Code | Frontier | 4 | 0 | 3 | 2 | 100% |
| DeepSeek V4 Flash | Mid | 2 | 2 | 1 | 3 | 50% |
| Prompt | Tier | Cursor | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-revenue-ops-copilot | Advanced | 195 | 67 | 7 | 139 | 74% |
| ai-revenue-ops-copilot | Intermediate | 209 | 39 | 2 | 170 | 84% |
| ai-support-agent-platform | Advanced | 207 | 40 | 0 | 182 | 84% |
| ai-support-agent-platform | Intermediate | 67 | 82 | 211 | 74 | 45% |
| ai-revenue-ops-copilot | Beginner | 87 | 55 | 231 | 62 | 61% |
| ai-support-agent-platform | Beginner | 69 | 40 | 289 | 39 | 63% |
| ai-engineering-workflow | Intermediate | 16 | 0 | 0 | 3 | 100% |
| ai-engineering-workflow | Beginner | 15 | 0 | 0 | 4 | 100% |
| ai-engineering-workflow | Advanced | 10 | 0 | 0 | 5 | 100% |
