| Metric | Value |
|---|---|
| LangGraph wins | 214 |
| LangChain wins | 132 |
| Abstains (no tool) | 1 |
| Other tool chosen | 9 |
| Decisive cases | 346 |
| LangGraph win rate (unweighted) | 61.8% |
| 95% CI | 56.6% - 66.8% |
| LangGraph win rate (weighted) | 61.8% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | LangGraph | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| Claude Opus 4.8 | Frontier | 18 | 0 | 0 | 0 | 100% |
| Gemini 3.5 Flash | Small | 18 | 0 | 0 | 0 | 100% |
| GPT 5.3 Codex | Frontier | 18 | 0 | 0 | 0 | 100% |
| GPT 5.4 Mini | Mid | 18 | 0 | 0 | 0 | 100% |
| GPT 5.5 | Frontier | 18 | 0 | 0 | 0 | 100% |
| Kimi K2.7 Code | Frontier | 18 | 0 | 0 | 0 | 100% |
| Claude Haiku 4.5 | Small | 17 | 1 | 0 | 0 | 94% |
| MiniMax M3 | Frontier | 15 | 3 | 0 | 0 | 83% |
| Mistral Small 4 | Mid | 15 | 3 | 0 | 0 | 83% |
| Claude Sonnet 4.6 | Frontier | 12 | 6 | 0 | 0 | 67% |
| Qwen3 Coder Next | Mid | 4 | 14 | 0 | 0 | 22% |
| Gemini 2.5 Pro | Frontier | 0 | 18 | 0 | 0 | 0% |
| Llama 4 Maverick | Frontier | 0 | 18 | 0 | 0 | 0% |
| GLM 5.2 | Frontier | 16 | 1 | 0 | 1 | 94% |
| DeepSeek V4 Flash | Mid | 7 | 10 | 0 | 0 | 41% |
| DeepSeek V4 Pro | Frontier | 5 | 12 | 0 | 1 | 29% |
| DeepSeek R1 0528 | Frontier | 0 | 17 | 1 | 0 | 0% |
| MiMo V2.5 Pro | Frontier | 12 | 4 | 0 | 2 | 75% |
| Devstral 2 2512 | Mid | 3 | 12 | 0 | 0 | 20% |
| Llama 4 Scout | Small | 0 | 13 | 0 | 5 | 0% |
| Prompt | Tier | LangGraph | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-agent-application | Advanced | 82 | 36 | 0 | 0 | 69% |
| ai-agent-application | Intermediate | 74 | 42 | 1 | 2 | 64% |
| ai-agent-application | Beginner | 58 | 54 | 0 | 7 | 52% |
