Design Arena — agenticslides(python-pptx)
Caveat: Head-to-head human votes on generated designs/code, mirrored via OpenRouter.
| # | Model | Score | |
|---|---|---|---|
| 1 | Claude Opus 4.7Anthropic | 1331 | |
| 2 | Claude Opus 4.8Anthropic | 1310 | |
| 3 | Gemini 3.5 FlashGoogle | 1242 | |
| 4 | GLM-5.1Zhipu AI | 1240 | |
| 5 | Kimi K2.6Moonshot AI | 1186 | |
| 6 | GLM-5V-TurboZhipu AI | 1183 | |
| 7 | GPT-5.5OpenAI | 1155 | |
| 8 | Gemini 3.1 Pro PreviewGoogle | 1107 | |
| 9 | Gemini 3 Flash PreviewGoogle | 1075 | |
| 10 | Grok 4.3xAI | 1068 |
Dot = elo on a 1054–1345 scale (position, not length). Scores link to model pages; every number keeps its source on the model page.