| Metric | Value |
|---|---|
| Langfuse wins | 1254 |
| LangSmith wins | 2329 |
| Abstains (no tool) | 146 |
| Other tool chosen | 625 |
| Decisive cases | 3583 |
| Langfuse win rate (unweighted) | 35.0% |
| 95% CI | 33.5% - 36.6% |
| Langfuse win rate (weighted) | 35.0% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model range | Tier | Langfuse | LangSmith | None | Other | A rate |
|---|---|---|---|---|---|---|
| DeepSeek Flash 3.2–4.1 | Mid | 38 | 179 | 4 | 6 | 18% |
| DeepSeek Pro 4–4 0813 | Frontier | 27 | 66 | 1 | 2 | 29% |
| DeepSeek R1 | Frontier | 0 | 205 | 2 | 21 | 0% |
| Devstral | Mid | 0 | 36 | 28 | 133 | 0% |
| Gemini Flash 2.5–3.8 | Small | 18 | 77 | 3 | 130 | 19% |
| Gemini Pro 2.5–3.1 Preview | Frontier | 8 | 207 | 8 | 5 | 4% |
| GLM 5 Turbo–5.3 | Frontier | 49 | 179 | 0 | 0 | 21% |
| GPT 5.4–6 Astra | Frontier | 169 | 54 | 4 | 1 | 76% |
| GPT Coding 5.3 Codex–6.1 Sol | Frontier | 200 | 27 | 1 | 0 | 88% |
| GPT Mini 5.4 Mini–6 Luna | Mid | 56 | 163 | 5 | 4 | 26% |
| Haiku | Small | 145 | 57 | 3 | 19 | 72% |
| Kimi 2.5–3 | Frontier | 98 | 111 | 4 | 1 | 47% |
| Llama Maverick | Frontier | 22 | 121 | 16 | 67 | 15% |
| Llama Scout | Small | 37 | 0 | 33 | 144 | 100% |
| MiMo Pro 2–2.6 | Frontier | 16 | 198 | 8 | 6 | 7% |
| MiniMax M 2.7–3 | Frontier | 51 | 148 | 5 | 22 | 26% |
| Mistral Small | Mid | 11 | 174 | 3 | 35 | 6% |
| Opus 4.6–5.5 | Frontier | 156 | 64 | 8 | 0 | 71% |
| Qwen Coder | Mid | 85 | 111 | 2 | 29 | 43% |
| Sonnet 4.6–5.5 | Frontier | 68 | 152 | 8 | 0 | 31% |
| Prompt | Tier | Langfuse | LangSmith | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Intermediate | 112 | 379 | 1 | 82 | 23% |
| ai-revenue-ops-copilot | Intermediate | 127 | 353 | 1 | 78 | 26% |
| ai-support-agent-platform | Beginner | 217 | 256 | 17 | 80 | 46% |
| ai-revenue-ops-copilot | Advanced | 135 | 332 | 3 | 95 | 29% |
| ai-support-agent-platform | Advanced | 170 | 293 | 1 | 108 | 37% |
| ai-revenue-ops-copilot | Beginner | 148 | 305 | 33 | 89 | 33% |
| ai-agent-application | Advanced | 56 | 89 | 0 | 11 | 39% |
| ai-agent-application | Intermediate | 34 | 108 | 1 | 12 | 24% |
| ai-engineering-workflow | Intermediate | 87 | 45 | 0 | 25 | 66% |
| ai-agent-application | Beginner | 29 | 101 | 24 | 4 | 22% |
| ai-engineering-workflow | Advanced | 81 | 43 | 0 | 31 | 65% |
| ai-engineering-workflow | Beginner | 58 | 25 | 65 | 10 | 70% |
