CapArena-Auto vs CogVLM-19B: leaderboard
CapArena-Auto vs CogVLM-19B measures pairwise caption wins and losses against the CogVLM-19B baseline.
Metric: CapArena-Auto Score (vs CogVLM-19B). Source: huggingface.co. Saturation forecast: Estimated already saturated. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 1.5 Pro (002) | 61 | #349 |
| 2 | GPT-4o (2024-08-06) | 55.5 | #336 |
| 3 | Qwen 2.5 VL 72B Instruct | 49 | #377 |
| 4 | Gemini 2.0 Flash (Preview) | 39.5 | #370 |
| 5 | Claude 3.5 Sonnet (20240620) | 30 | #299 |
| 6 | GPT-4o Mini (2024-07-18) | 17 | #558 |
| 7 | Llama 3.2 90B Vision Instruct | -13 | #622 |
| 8 | MiniCPM-V-2.6 | -34 | #836 |
| 9 | Qwen 2 VL 7B Instruct | -59 | #826 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=caparena-auto-vs-cogvlm-19b · How It Works · Data refreshed daily, snapshot 2026-10-20.