Leading: LangGraph (64.5%)
| Metric | Value |
|---|---|
| Anthropic wins | 146 |
| LangGraph wins | 265 |
| Abstains (no tool) | 103 |
| Other tool chosen | 3004 |
| Decisive cases | 411 |
| Anthropic win rate (unweighted) | 35.5% |
| 95% CI | 31.0% - 40.3% |
| Anthropic win rate (weighted) | 35.5% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Anthropic | LangGraph | None | Other | A rate |
|---|---|---|---|---|---|---|
| Claude Haiku 4.5 | Small | 81 | 0 | 0 | 107 | 100% |
| Kimi K2.5 | Frontier | 1 | 57 | 1 | 54 | 2% |
| GLM 5 Turbo | Frontier | 1 | 54 | 2 | 68 | 2% |
| Claude Sonnet 4.6 | Frontier | 0 | 47 | 0 | 142 | 0% |
| Mistral Small 4 | Mid | 9 | 34 | 14 | 95 | 21% |
| Gemini 3.5 Flash | Small | 3 | 27 | 1 | 30 | 10% |
| Kimi K2.7 Code | Frontier | 10 | 19 | 2 | 30 | 34% |
| MiniMax M2.7 | Frontier | 17 | 1 | 0 | 96 | 94% |
| GLM 5.2 | Frontier | 2 | 11 | 1 | 49 | 15% |
| MiniMax M3 | Frontier | 9 | 2 | 0 | 51 | 82% |
| GPT 5.5 | Frontier | 5 | 0 | 0 | 52 | 100% |
| Qwen3 Coder Next | Mid | 2 | 3 | 1 | 180 | 40% |
| MiMo V2.5 Pro | Frontier | 1 | 3 | 4 | 54 | 25% |
| Claude Opus 4.6 | Frontier | 0 | 4 | 0 | 122 | 0% |
| GPT 5.3 Codex | Frontier | 3 | 0 | 0 | 185 | 100% |
| DeepSeek V4 Flash | Mid | 1 | 0 | 8 | 50 | 100% |
| GPT 5.4 Mini | Mid | 1 | 0 | 1 | 185 | 100% |
| Claude Opus 4.8 | Frontier | 0 | 1 | 0 | 61 | 0% |
| Devstral 2 2512 | Mid | 0 | 1 | 12 | 149 | 0% |
| GPT 5.4 | Frontier | 0 | 1 | 0 | 125 | 0% |
| DeepSeek R1 0528 | Frontier | 0 | 0 | 18 | 168 | n/a |
| DeepSeek V3.2 | Mid | 0 | 0 | 0 | 125 | n/a |
| DeepSeek V4 Pro | Frontier | 0 | 0 | 9 | 54 | n/a |
| Gemini 2.5 Flash | Small | 0 | 0 | 0 | 126 | n/a |
| Gemini 2.5 Pro | Frontier | 0 | 0 | 1 | 187 | n/a |
| Llama 4 Maverick | Frontier | 0 | 0 | 21 | 163 | n/a |
| Llama 4 Scout | Small | 0 | 0 | 7 | 172 | n/a |
| MiMo V2 Pro | Frontier | 0 | 0 | 0 | 124 | n/a |
| Prompt | Tier | Anthropic | LangGraph | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-revenue-ops-copilot | Intermediate | 32 | 99 | 1 | 397 | 24% |
| ai-support-agent-platform | Intermediate | 28 | 43 | 0 | 460 | 39% |
| ai-revenue-ops-copilot | Advanced | 5 | 63 | 3 | 443 | 7% |
| ai-support-agent-platform | Advanced | 13 | 43 | 1 | 472 | 23% |
| ai-revenue-ops-copilot | Beginner | 20 | 6 | 0 | 476 | 77% |
| ai-agent-application | Advanced | 15 | 6 | 22 | 92 | 71% |
| ai-support-agent-platform | Beginner | 11 | 5 | 1 | 491 | 69% |
| ai-agent-application | Beginner | 12 | 0 | 39 | 87 | 100% |
| ai-agent-application | Intermediate | 10 | 0 | 36 | 86 | 100% |
