| Metric | Value |
|---|---|
| Ragas wins | 102 |
| Arize AI wins | 105 |
| Abstains (no tool) | 90 |
| Other tool chosen | 2147 |
| Decisive cases | 207 |
| Ragas win rate (unweighted) | 49.3% |
| 95% CI | 42.5% - 56.0% |
| Ragas win rate (weighted) | 49.3% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Ragas | Arize AI | None | Other | A rate |
|---|---|---|---|---|---|---|
| Llama 4 Scout | Small | 3 | 40 | 4 | 74 | 7% |
| Gemini 2.5 Flash | Small | 0 | 27 | 1 | 99 | 0% |
| MiniMax M2.7 | Frontier | 24 | 1 | 5 | 99 | 96% |
| MiMo V2 Pro | Frontier | 24 | 0 | 8 | 100 | 100% |
| Claude Opus 4.6 | Frontier | 18 | 0 | 0 | 114 | 100% |
| Llama 4 Maverick | Frontier | 0 | 13 | 0 | 114 | 0% |
| Mistral Small 4 | Mid | 9 | 2 | 1 | 112 | 82% |
| Gemini 2.5 Pro | Frontier | 0 | 10 | 9 | 113 | 0% |
| Devstral 2 2512 | Mid | 1 | 8 | 4 | 112 | 11% |
| GPT 5.4 Mini | Mid | 7 | 0 | 3 | 122 | 100% |
| Claude Sonnet 4.6 | Frontier | 6 | 0 | 0 | 126 | 100% |
| GLM 5 Turbo | Frontier | 6 | 0 | 19 | 107 | 100% |
| DeepSeek R1 0528 | Frontier | 0 | 3 | 7 | 122 | 0% |
| DeepSeek V3.2 | Mid | 2 | 0 | 22 | 104 | 100% |
| GPT 5.4 | Frontier | 1 | 0 | 0 | 131 | 100% |
| Kimi K2.5 | Frontier | 1 | 0 | 3 | 115 | 100% |
| Qwen3 Coder Next | Mid | 0 | 1 | 3 | 127 | 0% |
| Claude Haiku 4.5 | Small | 0 | 0 | 1 | 124 | n/a |
| GPT 5.3 Codex | Frontier | 0 | 0 | 0 | 132 | n/a |
| Prompt | Tier | Ragas | Arize AI | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Advanced | 36 | 18 | 5 | 351 | 67% |
| ai-support-agent-platform | Beginner | 34 | 18 | 64 | 295 | 65% |
| ai-revenue-ops-copilot | Advanced | 9 | 23 | 2 | 366 | 28% |
| ai-revenue-ops-copilot | Beginner | 13 | 15 | 10 | 372 | 46% |
| ai-support-agent-platform | Intermediate | 8 | 20 | 5 | 376 | 29% |
| ai-revenue-ops-copilot | Intermediate | 2 | 11 | 4 | 387 | 15% |