Appwrite Arena (Without Skills): leaderboard
Questions about Appwrite services across areas such as auth, databases, functions, storage and messaging; the August 2026 run used 191 questions, 165 multiple choice and 26 free form. Here no documentation is supplied and the model answers from training data. Multiple choice is auto scored and free form judged by another model, giving an overall percentage where higher is better. Source: github.com/appwrite/arena.
Metric: Overall Score (%). Source: raw.githubusercontent.com. Saturation forecast: Estimated already saturated. 24 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 | 97.7 | #16 |
| 2 | Claude Opus 4.8 | 97.4 | #37 |
| 3 | Claude Opus 5 | 97.4 | #12 |
| 4 | Claude Opus 4.7 | 96.2 | #49 |
| 5 | GPT-5.6 Sol | 95.5 | #19 |
| 6 | Kimi K3 | 95 | #30 |
| 7 | Muse Spark 1.2 | 94.8 | #62 |
| 8 | GPT-5.5 | 94.1 | #31 |
| 9 | Kimi K2.6 | 93.6 | #104 |
| 10 | Gemini 3.7 Flash | 93.6 | #23 |
| 11 | GPT-5.6 Terra | 93.5 | #41 |
| 12 | GPT-5.6 Luna | 92.9 | #68 |
| 13 | Grok 4.6 | 91.6 | #26 |
| 14 | Grok Build 0.1 | 91.5 | #117 |
| 15 | Gemini 3.1 Pro (Preview) | 91 | #58 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=appwrite-arena-without-skills · How It Works · Data refreshed daily, snapshot 2026-10-20.