| Metric | Value |
|---|---|
| Celery wins | 131 |
| BullMQ wins | 110 |
| Abstains (no tool) | 3 |
| Other tool chosen | 108 |
| Decisive cases | 241 |
| Celery win rate (unweighted) | 54.4% |
| 95% CI | 48.0% - 60.5% |
| Celery win rate (weighted) | 54.4% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Celery | BullMQ | None | Other | A rate |
|---|---|---|---|---|---|---|
| Claude Opus 4.8 | Frontier | 10 | 11 | 0 | 0 | 48% |
| Claude Sonnet 4.6 | Frontier | 9 | 11 | 0 | 0 | 45% |
| DeepSeek R1 0528 | Frontier | 15 | 4 | 0 | 1 | 79% |
| MiMo V2.5 Pro | Frontier | 14 | 5 | 0 | 2 | 74% |
| Gemini 3.5 Flash | Small | 7 | 9 | 0 | 5 | 44% |
| Claude Haiku 4.5 | Small | 13 | 1 | 0 | 7 | 93% |
| DeepSeek V4 Pro | Frontier | 8 | 6 | 0 | 7 | 57% |
| GPT 5.3 Codex | Frontier | 10 | 3 | 0 | 8 | 77% |
| Gemini 2.5 Pro | Frontier | 5 | 8 | 2 | 4 | 38% |
| DeepSeek V4 Flash | Mid | 8 | 3 | 0 | 2 | 73% |
| Devstral 2 2512 | Mid | 6 | 5 | 0 | 1 | 55% |
| GPT 5.4 Mini | Mid | 3 | 8 | 1 | 5 | 27% |
| Qwen3 Coder Next | Mid | 0 | 10 | 0 | 4 | 0% |
| Llama 4 Scout | Small | 9 | 0 | 0 | 2 | 100% |
| Llama 4 Maverick | Frontier | 6 | 3 | 0 | 0 | 67% |
| GLM 5.2 | Frontier | 1 | 7 | 0 | 13 | 13% |
| Kimi K2.7 Code | Frontier | 1 | 7 | 0 | 11 | 13% |
| GPT 5.5 | Frontier | 4 | 3 | 0 | 9 | 57% |
| Mistral Small 4 | Mid | 0 | 5 | 0 | 9 | 0% |
| MiniMax M3 | Frontier | 2 | 1 | 0 | 18 | 67% |
| Prompt | Tier | Celery | BullMQ | None | Other | A rate |
|---|---|---|---|---|---|---|
| full-stack-web-app | Intermediate | 39 | 59 | 1 | 21 | 40% |
| full-stack-web-app | Beginner | 76 | 9 | 0 | 37 | 89% |
| full-stack-web-app | Advanced | 16 | 42 | 2 | 50 | 28% |
