| Metric | Value |
|---|---|
| Ably wins | 929 |
| Firebase wins | 744 |
| Abstains (no tool) | 42 |
| Other tool chosen | 2021 |
| Decisive cases | 1673 |
| Ably win rate (unweighted) | 55.5% |
| 95% CI | 53.1% - 57.9% |
| Ably win rate (weighted) | 55.5% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Ably | Firebase | None | Other | A rate |
|---|---|---|---|---|---|---|
| Devstral 2 2512 | Mid | 0 | 195 | 0 | 3 | 0% |
| Llama 4 Scout | Small | 62 | 60 | 0 | 73 | 51% |
| GPT 5.3 Codex | Frontier | 120 | 0 | 0 | 78 | 100% |
| Mistral Small 4 | Mid | 39 | 79 | 0 | 78 | 33% |
| GPT 5.4 Mini | Mid | 95 | 13 | 2 | 84 | 88% |
| DeepSeek R1 0528 | Frontier | 66 | 38 | 31 | 62 | 63% |
| Qwen3 Coder Next | Mid | 0 | 102 | 0 | 95 | 0% |
| GPT 5.4 | Frontier | 93 | 6 | 0 | 96 | 94% |
| Gemini 2.5 Flash | Small | 3 | 83 | 0 | 112 | 3% |
| Claude Sonnet 4.6 | Frontier | 77 | 0 | 0 | 121 | 100% |
| Gemini 2.5 Pro | Frontier | 73 | 0 | 0 | 125 | 100% |
| MiniMax M2.7 | Frontier | 26 | 45 | 0 | 125 | 37% |
| Kimi K2.5 | Frontier | 70 | 0 | 5 | 121 | 100% |
| Llama 4 Maverick | Frontier | 70 | 0 | 0 | 128 | 100% |
| Claude Haiku 4.5 | Small | 0 | 67 | 0 | 125 | 0% |
| GLM 5 Turbo | Frontier | 65 | 0 | 3 | 130 | 100% |
| MiMo V2 Pro | Frontier | 9 | 54 | 0 | 134 | 14% |
| Claude Opus 4.6 | Frontier | 49 | 0 | 0 | 148 | 100% |
| DeepSeek V3.2 | Mid | 12 | 2 | 1 | 183 | 86% |
| Prompt | Tier | Ably | Firebase | None | Other | A rate |
|---|---|---|---|---|---|---|
| chat-application | Advanced | 253 | 38 | 1 | 122 | 87% |
| chat-application | Intermediate | 84 | 155 | 6 | 171 | 35% |
| social-media-platform | Advanced | 187 | 25 | 3 | 202 | 88% |
| project-management-tool | Advanced | 156 | 30 | 0 | 231 | 84% |
| social-media-platform | Intermediate | 95 | 80 | 4 | 236 | 54% |
| chat-application | Beginner | 31 | 127 | 3 | 254 | 20% |
| project-management-tool | Intermediate | 44 | 100 | 3 | 268 | 31% |
| project-management-tool | Beginner | 18 | 120 | 5 | 274 | 13% |
| social-media-platform | Beginner | 61 | 69 | 17 | 263 | 47% |
