| Metric | Value |
|---|---|
| Langfuse wins | 1119 |
| LangSmith wins | 2170 |
| Abstains (no tool) | 129 |
| Other tool chosen | 587 |
| Decisive cases | 3289 |
| Langfuse win rate (unweighted) | 34.0% |
| 95% CI | 32.4% - 35.7% |
| Langfuse win rate (weighted) | 34.0% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Langfuse | LangSmith | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.3 Codex | Frontier | 187 | 23 | 0 | 0 | 89% |
| Claude Sonnet 4.6 | Frontier | 54 | 149 | 7 | 0 | 27% |
| GPT 5.4 Mini | Mid | 45 | 156 | 5 | 4 | 22% |
| Gemini 2.5 Pro | Frontier | 7 | 192 | 6 | 5 | 4% |
| DeepSeek R1 0528 | Frontier | 0 | 188 | 2 | 20 | 0% |
| Claude Haiku 4.5 | Small | 132 | 54 | 3 | 17 | 71% |
| Qwen3 Coder Next | Mid | 77 | 103 | 2 | 27 | 43% |
| Mistral Small 4 | Mid | 10 | 159 | 3 | 33 | 6% |
| Llama 4 Maverick | Frontier | 19 | 113 | 14 | 63 | 14% |
| GPT 5.4 | Frontier | 114 | 12 | 0 | 0 | 90% |
| Claude Opus 4.6 | Frontier | 106 | 20 | 0 | 0 | 84% |
| GLM 5 Turbo | Frontier | 12 | 114 | 0 | 0 | 10% |
| DeepSeek V3.2 | Mid | 7 | 114 | 0 | 5 | 6% |
| MiMo V2 Pro | Frontier | 1 | 119 | 2 | 4 | 1% |
| Kimi K2.5 | Frontier | 48 | 62 | 4 | 0 | 44% |
| MiniMax M2.7 | Frontier | 1 | 100 | 3 | 20 | 1% |
| GLM 5.2 | Frontier | 33 | 51 | 0 | 0 | 39% |
| MiniMax M3 | Frontier | 47 | 35 | 1 | 1 | 57% |
| Kimi K2.7 Code | Frontier | 40 | 42 | 0 | 1 | 49% |
| DeepSeek V4 Pro | Frontier | 16 | 65 | 1 | 2 | 20% |
| GPT 5.5 | Frontier | 43 | 37 | 3 | 1 | 54% |
| DeepSeek V4 Flash | Mid | 28 | 51 | 4 | 0 | 35% |
| Gemini 3.5 Flash | Small | 9 | 70 | 1 | 4 | 11% |
| Claude Opus 4.8 | Frontier | 36 | 41 | 7 | 0 | 47% |
| MiMo V2.5 Pro | Frontier | 10 | 66 | 6 | 2 | 13% |
| Llama 4 Scout | Small | 36 | 0 | 29 | 133 | 100% |
| Devstral 2 2512 | Mid | 0 | 30 | 25 | 125 | 0% |
| Gemini 2.5 Flash | Small | 1 | 4 | 1 | 120 | 20% |
| Prompt | Tier | Langfuse | LangSmith | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Intermediate | 102 | 355 | 1 | 77 | 22% |
| ai-revenue-ops-copilot | Intermediate | 115 | 330 | 1 | 76 | 26% |
| ai-support-agent-platform | Beginner | 199 | 241 | 17 | 74 | 45% |
| ai-revenue-ops-copilot | Advanced | 123 | 313 | 2 | 89 | 28% |
| ai-support-agent-platform | Advanced | 155 | 275 | 1 | 102 | 36% |
| ai-revenue-ops-copilot | Beginner | 137 | 284 | 30 | 85 | 33% |
| ai-agent-application | Advanced | 47 | 79 | 0 | 10 | 37% |
| ai-agent-application | Intermediate | 26 | 98 | 1 | 12 | 21% |
| ai-engineering-workflow | Intermediate | 71 | 43 | 0 | 23 | 62% |
| ai-agent-application | Beginner | 25 | 89 | 21 | 3 | 22% |
| ai-engineering-workflow | Advanced | 68 | 40 | 0 | 27 | 63% |
| ai-engineering-workflow | Beginner | 51 | 23 | 55 | 9 | 69% |
