LMArena Vision Arena: leaderboard
Crowdsourced pairwise human-preference leaderboard for vision-language model responses in LMArena.
Metric: Arena rating. Source: lmarena.ai. Source: benchmarklist.com. Saturation forecast: Around March 2028. 25 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 | 1313.06 | #16 |
| 2 | Claude Opus 4.7 (High) | 1301.5 | #49 (Claude Opus 4.7) |
| 3 | Qwen 3.8 Max | 1299.71 | #35 |
| 4 | Claude Opus 4.7 | 1299.41 | #49 |
| 5 | Claude Opus 4.6 (High) | 1298.76 | #63 (Claude Opus 4.6) |
| 6 | Muse Spark | 1293.73 | #64 |
| 7 | Claude Opus 4.6 | 1293.13 | #63 |
| 8 | Muse Spark 1.2 (xHigh) | 1291.76 | #62 (Muse Spark 1.2) |
| 9 | Claude Opus 5 (High) | 1290.35 | #12 (Claude Opus 5) |
| 10 | Gemini 3 Pro | 1288.59 | #80 |
| 11 | GPT-5.5 | 1286.1 | #31 |
| 12 | Claude Opus 4.8 (High) | 1285.3 | #37 (Claude Opus 4.8) |
| 13 | Gemini 3.6 Flash (High) | 1284.89 | #51 (Gemini 3.6 Flash) |
| 14 | Gemini 3.5 Flash (High) | 1284.3 | #60 (Gemini 3.5 Flash) |
| 15 | GPT-5.5 (High) | 1283.57 | #31 (GPT-5.5) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=lmarena-vision-arena · How It Works · Data refreshed daily, snapshot 2026-10-20.