| Metric | Value |
|---|---|
| Langfuse wins | 50 |
| LangChain wins | 70 |
| Abstains (no tool) | 105 |
| Other tool chosen | 2450 |
| Decisive cases | 120 |
| Langfuse win rate (unweighted) | 41.7% |
| 95% CI | 33.2% - 50.6% |
| Langfuse win rate (weighted) | 41.7% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Langfuse | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| Qwen3 Coder Next | Mid | 17 | 17 | 3 | 104 | 50% |
| Gemini 2.5 Flash | Small | 0 | 30 | 1 | 96 | 0% |
| Claude Sonnet 4.6 | Frontier | 13 | 0 | 1 | 130 | 100% |
| Llama 4 Scout | Small | 4 | 6 | 7 | 114 | 40% |
| Llama 4 Maverick | Frontier | 0 | 10 | 2 | 125 | 0% |
| DeepSeek V3.2 | Mid | 1 | 7 | 22 | 98 | 13% |
| GPT 5.4 Mini | Mid | 7 | 0 | 3 | 133 | 100% |
| Claude Haiku 4.5 | Small | 4 | 0 | 1 | 132 | 100% |
| Claude Opus 4.8 | Frontier | 2 | 0 | 1 | 9 | 100% |
| DeepSeek V4 Flash | Mid | 1 | 0 | 1 | 10 | 100% |
| Mistral Small 4 | Mid | 1 | 0 | 2 | 133 | 100% |
| Claude Opus 4.6 | Frontier | 0 | 0 | 0 | 132 | n/a |
| DeepSeek R1 0528 | Frontier | 0 | 0 | 7 | 137 | n/a |
| DeepSeek V4 Pro | Frontier | 0 | 0 | 1 | 11 | n/a |
| Devstral 2 2512 | Mid | 0 | 0 | 6 | 129 | n/a |
| Gemini 2.5 Pro | Frontier | 0 | 0 | 9 | 135 | n/a |
| Gemini 3.5 Flash | Small | 0 | 0 | 1 | 11 | n/a |
| GLM 5 Turbo | Frontier | 0 | 0 | 19 | 113 | n/a |
| GLM 5.2 | Frontier | 0 | 0 | 0 | 12 | n/a |
| GPT 5.3 Codex | Frontier | 0 | 0 | 0 | 144 | n/a |
| GPT 5.4 | Frontier | 0 | 0 | 0 | 132 | n/a |
| GPT 5.5 | Frontier | 0 | 0 | 0 | 12 | n/a |
| Kimi K2.5 | Frontier | 0 | 0 | 3 | 116 | n/a |
| Kimi K2.7 Code | Frontier | 0 | 0 | 1 | 11 | n/a |
| MiMo V2 Pro | Frontier | 0 | 0 | 8 | 124 | n/a |
| MiMo V2.5 Pro | Frontier | 0 | 0 | 0 | 12 | n/a |
| MiniMax M2.7 | Frontier | 0 | 0 | 5 | 124 | n/a |
| MiniMax M3 | Frontier | 0 | 0 | 1 | 11 | n/a |
| Prompt | Tier | Langfuse | LangChain | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Beginner | 23 | 12 | 66 | 330 | 66% |
| ai-revenue-ops-copilot | Intermediate | 8 | 23 | 4 | 389 | 26% |
| ai-revenue-ops-copilot | Beginner | 4 | 19 | 10 | 397 | 17% |
| ai-revenue-ops-copilot | Advanced | 0 | 10 | 2 | 407 | 0% |
| ai-support-agent-platform | Intermediate | 7 | 2 | 5 | 414 | 78% |
| ai-support-agent-platform | Advanced | 5 | 4 | 5 | 416 | 56% |
| ai-engineering-workflow | Beginner | 2 | 0 | 8 | 10 | 100% |
| ai-engineering-workflow | Intermediate | 1 | 0 | 0 | 16 | 100% |
| ai-agent-application | Intermediate | 0 | 0 | 0 | 20 | n/a |
| ai-agent-application | Advanced | 0 | 0 | 0 | 18 | n/a |
| ai-agent-application | Beginner | 0 | 0 | 5 | 15 | n/a |
| ai-engineering-workflow | Advanced | 0 | 0 | 0 | 18 | n/a |
