| Metric | Value |
|---|---|
| OpenAI wins | 1525 |
| LangChain wins | 905 |
| Abstains (no tool) | 93 |
| Other tool chosen | 938 |
| Decisive cases | 2430 |
| OpenAI win rate (unweighted) | 62.8% |
| 95% CI | 60.8% - 64.7% |
| OpenAI win rate (weighted) | 62.8% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | OpenAI | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| Gemini 2.5 Pro | Frontier | 70 | 111 | 2 | 2 | 39% |
| GPT 5.3 Codex | Frontier | 180 | 0 | 0 | 5 | 100% |
| GPT 5.4 Mini | Mid | 173 | 0 | 1 | 10 | 100% |
| Qwen3 Coder Next | Mid | 80 | 87 | 1 | 15 | 48% |
| DeepSeek R1 0528 | Frontier | 155 | 11 | 15 | 2 | 93% |
| Llama 4 Maverick | Frontier | 0 | 147 | 18 | 15 | 0% |
| MiMo V2 Pro | Frontier | 115 | 13 | 0 | 2 | 90% |
| Claude Opus 4.6 | Frontier | 63 | 65 | 0 | 4 | 49% |
| Claude Sonnet 4.6 | Frontier | 70 | 53 | 0 | 63 | 57% |
| GPT 5.4 | Frontier | 122 | 0 | 0 | 10 | 100% |
| Llama 4 Scout | Small | 0 | 110 | 6 | 61 | 0% |
| Gemini 2.5 Flash | Small | 57 | 44 | 0 | 31 | 56% |
| DeepSeek V3.2 | Mid | 60 | 40 | 0 | 31 | 60% |
| Devstral 2 2512 | Mid | 11 | 67 | 12 | 78 | 14% |
| Mistral Small 4 | Mid | 11 | 53 | 12 | 73 | 17% |
| Kimi K2.5 | Frontier | 28 | 27 | 1 | 62 | 51% |
| Claude Opus 4.8 | Frontier | 51 | 1 | 0 | 1 | 98% |
| GPT 5.5 | Frontier | 43 | 0 | 0 | 6 | 100% |
| DeepSeek V4 Pro | Frontier | 40 | 3 | 9 | 2 | 93% |
| GLM 5 Turbo | Frontier | 7 | 34 | 2 | 88 | 17% |
| GLM 5.2 | Frontier | 34 | 4 | 1 | 15 | 89% |
| MiMo V2.5 Pro | Frontier | 28 | 10 | 4 | 11 | 74% |
| DeepSeek V4 Flash | Mid | 29 | 5 | 7 | 9 | 85% |
| MiniMax M3 | Frontier | 24 | 3 | 0 | 26 | 89% |
| MiniMax M2.7 | Frontier | 24 | 2 | 0 | 94 | 92% |
| Gemini 3.5 Flash | Small | 15 | 11 | 1 | 25 | 58% |
| Kimi K2.7 Code | Frontier | 19 | 1 | 1 | 31 | 95% |
| Claude Haiku 4.5 | Small | 16 | 3 | 0 | 166 | 84% |
| Prompt | Tier | OpenAI | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Advanced | 230 | 174 | 1 | 124 | 57% |
| ai-support-agent-platform | Intermediate | 223 | 174 | 0 | 134 | 56% |
| ai-revenue-ops-copilot | Intermediate | 167 | 205 | 1 | 156 | 45% |
| ai-revenue-ops-copilot | Beginner | 300 | 71 | 0 | 129 | 81% |
| ai-revenue-ops-copilot | Advanced | 158 | 208 | 4 | 144 | 43% |
| ai-support-agent-platform | Beginner | 272 | 50 | 1 | 187 | 84% |
| ai-agent-application | Advanced | 63 | 9 | 18 | 26 | 88% |
| ai-agent-application | Beginner | 60 | 9 | 35 | 15 | 87% |
| ai-agent-application | Intermediate | 52 | 5 | 33 | 23 | 91% |
