Chatbot Arena (Vision - Humor): leaderboard
The arena.ai vision board restricted to humour prompts, where users ask a model to pick out what is funny in an image or to explain the joke; one of the site's own examples asks for a better caption instead. Voters compare two anonymous answers to the same image and prompt. The rating is refit over the conversations in this category alone, and higher is better.
Metric: Arena Score. Source: arena.ai. 91 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 (High) | 1324 | #16 (Claude Fable 5) |
| 2 | Muse Spark | 1323 | #64 |
| 3 | Gemini 3 Pro | 1317 | #80 |
| 4 | Gemini 3.1 Pro (Preview) | 1300 | #58 |
| 5 | Gemini 3.7 Flash (High) | 1299 | #23 (Gemini 3.7 Flash) |
| 6 | Claude Opus 4.6 (High) | 1296 | #63 (Claude Opus 4.6) |
| 7 | Gemini 3.5 Flash (Medium) | 1295 | #60 (Gemini 3.5 Flash) |
| 8 | Claude Opus 4.6 | 1294 | #63 |
| 9 | Qwen 3.8 Max | 1291 | #35 |
| 10 | Claude Opus 4.7 (High) | 1289 | #49 (Claude Opus 4.7) |
| 11 | Gemini 3.5 Flash (High) | 1288 | #60 (Gemini 3.5 Flash) |
| 12 | GPT-5.5 Instant | 1287 | #129 |
| 13 | Claude Opus 4.7 | 1286 | #49 |
| 14 | Muse Spark 1.3 (xHigh) | 1283 | #24 (Muse Spark 1.3) |
| 15 | GPT-5.5 (High) | 1282 | #31 (GPT-5.5) |
Interactive version: theaggregate.ai/benchmark?slug=chatbot-arena-vision-humor · How It Works · Data refreshed daily, snapshot 2026-10-20.