| Metric | Value |
|---|---|
| LangChain wins | 181 |
| LangGraph wins | 339 |
| Abstains (no tool) | 1 |
| Other tool chosen | 12 |
| Decisive cases | 520 |
| LangChain win rate (unweighted) | 34.8% |
| 95% CI | 30.8% - 39.0% |
| LangChain win rate (weighted) | 34.8% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model range | Tier | LangChain | LangGraph | None | Other | A rate |
|---|---|---|---|---|---|---|
| DeepSeek Flash 3.2–4.1 | Mid | 13 | 13 | 0 | 0 | 50% |
| DeepSeek Pro 4–4 0813 | Frontier | 13 | 13 | 0 | 1 | 50% |
| DeepSeek R1 | Frontier | 26 | 0 | 1 | 0 | 100% |
| Devstral | Mid | 18 | 3 | 0 | 0 | 86% |
| Gemini Flash 2.5–3.8 | Small | 0 | 27 | 0 | 0 | 0% |
| Gemini Pro 2.5–3.1 Preview | Frontier | 22 | 5 | 0 | 0 | 81% |
| GLM 5 Turbo–5.3 | Frontier | 1 | 25 | 0 | 1 | 4% |
| GPT 5.4–6 Astra | Frontier | 0 | 27 | 0 | 0 | 0% |
| GPT Coding 5.3 Codex–6.1 Sol | Frontier | 0 | 27 | 0 | 0 | 0% |
| GPT Mini 5.4 Mini–6 Luna | Mid | 0 | 27 | 0 | 0 | 0% |
| Haiku | Small | 1 | 26 | 0 | 0 | 4% |
| Kimi 2.5–3 | Frontier | 0 | 27 | 0 | 0 | 0% |
| Llama Maverick | Frontier | 26 | 1 | 0 | 0 | 96% |
| Llama Scout | Small | 19 | 0 | 0 | 8 | 100% |
| MiMo Pro 2–2.6 | Frontier | 5 | 20 | 0 | 2 | 20% |
| MiniMax M 2.7–3 | Frontier | 4 | 23 | 0 | 0 | 15% |
| Mistral Small | Mid | 5 | 22 | 0 | 0 | 19% |
| Opus 4.6–5.5 | Frontier | 0 | 27 | 0 | 0 | 0% |
| Qwen Coder | Mid | 21 | 6 | 0 | 0 | 78% |
| Sonnet 4.6–5.5 | Frontier | 7 | 20 | 0 | 0 | 26% |
| Prompt | Tier | LangChain | LangGraph | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-agent-application | Advanced | 47 | 129 | 0 | 1 | 27% |
| ai-agent-application | Intermediate | 59 | 116 | 1 | 2 | 34% |
| ai-agent-application | Beginner | 75 | 94 | 0 | 9 | 44% |
