| Metric | Value |
|---|---|
| LangChain wins | 151 |
| LangGraph wins | 252 |
| Abstains (no tool) | 1 |
| Other tool chosen | 9 |
| Decisive cases | 403 |
| LangChain win rate (unweighted) | 37.5% |
| 95% CI | 32.9% - 42.3% |
| LangChain win rate (weighted) | 37.5% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | LangChain | LangGraph | None | Other | A rate |
|---|---|---|---|---|---|---|
| Gemini 2.5 Pro | Frontier | 21 | 0 | 0 | 0 | 100% |
| Llama 4 Maverick | Frontier | 21 | 0 | 0 | 0 | 100% |
| Qwen3 Coder Next | Mid | 17 | 4 | 0 | 0 | 81% |
| Claude Sonnet 4.6 | Frontier | 7 | 14 | 0 | 0 | 33% |
| MiniMax M3 | Frontier | 4 | 17 | 0 | 0 | 19% |
| Mistral Small 4 | Mid | 3 | 18 | 0 | 0 | 14% |
| Claude Haiku 4.5 | Small | 1 | 20 | 0 | 0 | 5% |
| Claude Opus 4.8 | Frontier | 0 | 21 | 0 | 0 | 0% |
| Gemini 3.5 Flash | Small | 0 | 21 | 0 | 0 | 0% |
| GPT 5.3 Codex | Frontier | 0 | 21 | 0 | 0 | 0% |
| GPT 5.4 Mini | Mid | 0 | 21 | 0 | 0 | 0% |
| GPT 5.5 | Frontier | 0 | 21 | 0 | 0 | 0% |
| Kimi K2.7 Code | Frontier | 0 | 21 | 0 | 0 | 0% |
| DeepSeek R1 0528 | Frontier | 20 | 0 | 1 | 0 | 100% |
| DeepSeek V4 Pro | Frontier | 12 | 8 | 0 | 1 | 60% |
| DeepSeek V4 Flash | Mid | 11 | 9 | 0 | 0 | 55% |
| GLM 5.2 | Frontier | 1 | 19 | 0 | 1 | 5% |
| MiMo V2.5 Pro | Frontier | 5 | 14 | 0 | 2 | 26% |
| Llama 4 Scout | Small | 16 | 0 | 0 | 5 | 100% |
| Devstral 2 2512 | Mid | 12 | 3 | 0 | 0 | 80% |
| Prompt | Tier | LangChain | LangGraph | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-agent-application | Advanced | 41 | 96 | 0 | 0 | 30% |
| ai-agent-application | Intermediate | 48 | 87 | 1 | 2 | 36% |
| ai-agent-application | Beginner | 62 | 69 | 0 | 7 | 47% |
