| Metric | Value |
|---|---|
| OpenAI wins | 414 |
| LangChain wins | 283 |
| Abstains (no tool) | 4 |
| Other tool chosen | 296 |
| Decisive cases | 697 |
| OpenAI win rate (unweighted) | 59.4% |
| 95% CI | 55.7% - 63.0% |
| OpenAI win rate (weighted) | 59.4% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | OpenAI | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.3 Codex | Frontier | 54 | 0 | 0 | 0 | 100% |
| Claude Opus 4.6 | Frontier | 27 | 27 | 0 | 0 | 50% |
| MiMo V2 Pro | Frontier | 44 | 9 | 0 | 0 | 83% |
| Gemini 2.5 Pro | Frontier | 16 | 37 | 1 | 0 | 30% |
| GPT 5.4 Mini | Mid | 51 | 0 | 0 | 2 | 100% |
| DeepSeek R1 0528 | Frontier | 47 | 4 | 0 | 2 | 92% |
| GPT 5.4 | Frontier | 49 | 0 | 0 | 5 | 100% |
| Qwen3 Coder Next | Mid | 27 | 21 | 0 | 6 | 56% |
| Llama 4 Maverick | Frontier | 0 | 45 | 0 | 8 | 0% |
| DeepSeek V3.2 | Mid | 24 | 15 | 0 | 14 | 62% |
| Gemini 2.5 Flash | Small | 20 | 19 | 0 | 15 | 51% |
| Claude Sonnet 4.6 | Frontier | 18 | 19 | 0 | 17 | 49% |
| Llama 4 Scout | Small | 0 | 27 | 0 | 26 | 0% |
| Devstral 2 2512 | Mid | 2 | 22 | 0 | 28 | 8% |
| Kimi K2.5 | Frontier | 13 | 8 | 1 | 25 | 62% |
| GLM 5 Turbo | Frontier | 2 | 16 | 2 | 34 | 11% |
| Mistral Small 4 | Mid | 4 | 12 | 0 | 26 | 25% |
| MiniMax M2.7 | Frontier | 11 | 0 | 0 | 41 | 100% |
| Claude Haiku 4.5 | Small | 5 | 2 | 0 | 47 | 71% |
| Prompt | Tier | OpenAI | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Intermediate | 69 | 58 | 0 | 44 | 54% |
| ai-support-agent-platform | Advanced | 70 | 55 | 1 | 44 | 56% |
| ai-revenue-ops-copilot | Intermediate | 46 | 73 | 0 | 48 | 39% |
| ai-revenue-ops-copilot | Advanced | 47 | 71 | 2 | 44 | 40% |
| ai-revenue-ops-copilot | Beginner | 96 | 14 | 0 | 50 | 87% |
| ai-support-agent-platform | Beginner | 86 | 12 | 1 | 66 | 88% |
