Leading: Portkey (63.2%)
| Metric | Value |
|---|---|
| OpenRouter wins | 133 |
| Portkey wins | 228 |
| Abstains (no tool) | 7 |
| Other tool chosen | 384 |
| Decisive cases | 361 |
| OpenRouter win rate (unweighted) | 36.8% |
| 95% CI | 32.0% - 41.9% |
| OpenRouter win rate (weighted) | 36.8% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | OpenRouter | Portkey | None | Other | A rate |
|---|---|---|---|---|---|---|
| Gemini 2.5 Pro | Frontier | 0 | 41 | 0 | 1 | 0% |
| GPT 5.3 Codex | Frontier | 13 | 22 | 0 | 6 | 37% |
| MiniMax M3 | Frontier | 7 | 28 | 0 | 5 | 20% |
| GLM 5.2 | Frontier | 14 | 19 | 0 | 9 | 42% |
| DeepSeek R1 0528 | Frontier | 3 | 30 | 1 | 8 | 9% |
| DeepSeek V4 Pro | Frontier | 3 | 23 | 0 | 16 | 12% |
| DeepSeek V4 Flash | Mid | 3 | 19 | 0 | 20 | 14% |
| Qwen3 Coder Next | Mid | 20 | 0 | 0 | 17 | 100% |
| Claude Opus 4.8 | Frontier | 18 | 2 | 0 | 22 | 90% |
| Gemini 3.5 Flash | Small | 6 | 11 | 0 | 25 | 35% |
| Devstral 2 2512 | Mid | 0 | 15 | 0 | 13 | 0% |
| GPT 5.5 | Frontier | 10 | 4 | 0 | 24 | 71% |
| MiMo V2.5 Pro | Frontier | 9 | 5 | 0 | 27 | 64% |
| GPT 5.4 Mini | Mid | 13 | 0 | 1 | 28 | 100% |
| Claude Sonnet 4.6 | Frontier | 7 | 2 | 0 | 33 | 78% |
| Llama 4 Maverick | Frontier | 0 | 5 | 1 | 36 | 0% |
| Claude Haiku 4.5 | Small | 4 | 0 | 0 | 27 | 100% |
| Kimi K2.7 Code | Frontier | 3 | 1 | 0 | 26 | 75% |
| Llama 4 Scout | Small | 0 | 1 | 0 | 15 | 0% |
| Mistral Small 4 | Mid | 0 | 0 | 4 | 26 | n/a |
| Prompt | Tier | OpenRouter | Portkey | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-engineering-workflow | Beginner | 74 | 27 | 5 | 23 | 73% |
| ai-engineering-workflow | Advanced | 1 | 62 | 0 | 54 | 2% |
| ai-agent-application | Intermediate | 25 | 27 | 1 | 76 | 48% |
| ai-agent-application | Advanced | 4 | 48 | 0 | 73 | 8% |
| ai-engineering-workflow | Intermediate | 2 | 49 | 0 | 71 | 4% |
| ai-agent-application | Beginner | 27 | 15 | 1 | 87 | 64% |
