| Metric | Value |
|---|---|
| Langfuse wins | 646 |
| LangSmith wins | 1341 |
| Abstains (no tool) | 45 |
| Other tool chosen | 441 |
| Decisive cases | 1987 |
| Langfuse win rate (unweighted) | 32.5% |
| 95% CI | 30.5% - 34.6% |
| Langfuse win rate (weighted) | 32.5% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Langfuse | LangSmith | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.4 | Frontier | 119 | 13 | 0 | 0 | 90% |
| GPT 5.3 Codex | Frontier | 118 | 14 | 0 | 0 | 89% |
| Claude Opus 4.6 | Frontier | 110 | 22 | 0 | 0 | 83% |
| Claude Sonnet 4.6 | Frontier | 30 | 102 | 0 | 0 | 23% |
| GLM 5 Turbo | Frontier | 12 | 120 | 0 | 0 | 9% |
| GPT 5.4 Mini | Mid | 23 | 105 | 1 | 3 | 18% |
| DeepSeek V3.2 | Mid | 7 | 120 | 0 | 5 | 6% |
| MiMo V2 Pro | Frontier | 1 | 125 | 2 | 4 | 1% |
| Gemini 2.5 Pro | Frontier | 1 | 122 | 6 | 3 | 1% |
| DeepSeek R1 0528 | Frontier | 0 | 118 | 1 | 13 | 0% |
| Claude Haiku 4.5 | Small | 79 | 37 | 0 | 13 | 68% |
| Qwen3 Coder Next | Mid | 55 | 61 | 1 | 14 | 47% |
| Mistral Small 4 | Mid | 4 | 112 | 0 | 13 | 3% |
| Kimi K2.5 | Frontier | 50 | 65 | 4 | 0 | 43% |
| MiniMax M2.7 | Frontier | 1 | 105 | 3 | 21 | 1% |
| Llama 4 Maverick | Frontier | 11 | 75 | 0 | 46 | 13% |
| Llama 4 Scout | Small | 24 | 0 | 11 | 90 | 100% |
| Devstral 2 2512 | Mid | 0 | 21 | 15 | 90 | 0% |
| Gemini 2.5 Flash | Small | 1 | 4 | 1 | 126 | 20% |
| Prompt | Tier | Langfuse | LangSmith | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Intermediate | 81 | 264 | 1 | 70 | 23% |
| ai-revenue-ops-copilot | Intermediate | 87 | 251 | 1 | 64 | 26% |
| ai-support-agent-platform | Beginner | 159 | 174 | 11 | 69 | 48% |
| ai-revenue-ops-copilot | Advanced | 96 | 232 | 2 | 79 | 29% |
| ai-support-agent-platform | Advanced | 118 | 206 | 1 | 90 | 36% |
| ai-revenue-ops-copilot | Beginner | 105 | 214 | 29 | 69 | 33% |
