LMArena Vision arena · published July 12, 2026
Anthropic leads the LMArena vision arena.
As of July 12, 2026, Anthropic leads the LMArena vision arena with claude-fable-5 at 1335.1 across 4,383 battles, according to LMArena.
Current through July 12, 2026. 140 published snapshots. 113 of 143 models resolved (30 held for exact-match review).
| Rank | LMArena rating | Organization | Model | Reported interval | Gap to leader | Battles |
|---|---|---|---|---|---|---|
| 01 | 1335.1 | Anthropic | Claude Fable 5 | 1324.5 to 1345.8 | Leader | 4,383 |
| 02 | 1318.4 | Anthropic | claude-opus-4-7-thinking | 1311.4 to 1325.5 | -16.7 | 18,567 |
| 03 | 1316.5 | Anthropic | claude-opus-4-6 | 1309.8 to 1323.2 | -18.6 | 22,778 |
| 04 | 1315.5 | Anthropic | Claude Opus 4.7 | 1308.5 to 1322.5 | -19.6 | 19,049 |
| 05 | 1315.0 | Anthropic | claude-opus-4-6-thinking | 1308.0 to 1322.0 | -20.1 | 18,430 |
| 06 | 1308.8 | gemini-3.5-flash-medium | 1297.1 to 1320.5 | -26.3 | 3,350 | |
| 07 | 1307.3 | Meta | Muse Spark | 1297.8 to 1316.8 | -27.8 | 5,621 |
| 08 | 1304.9 | Gemini 3 Pro | 1297.2 to 1312.5 | -30.2 | 13,117 | |
| 09 | 1300.5 | gemini-3.5-flash-high | 1288.2 to 1312.8 | -34.6 | 2,988 | |
| 10 | 1298.8 | OpenAI | gpt-5.4-high | 1291.8 to 1305.9 | -36.3 | 18,422 |
| 11 | 1296.7 | OpenAI | gpt-5.5-high | 1289.3 to 1304.0 | -38.4 | 15,445 |
| 12 | 1295.5 | gemini-3.1-pro-preview | 1289.6 to 1301.5 | -39.6 | 34,899 | |
| 13 | 1295.4 | OpenAI | gpt-5.5 | 1288.1 to 1302.7 | -39.7 | 16,562 |
| 14 | 1294.8 | Anthropic | claude-opus-4-8-thinking | 1286.3 to 1303.3 | -40.3 | 9,124 |
| 15 | 1291.3 | Anthropic | claude-opus-4-8 | 1282.7 to 1299.8 | -43.8 | 9,505 |
| 16 | 1290.6 | OpenAI | gpt-5.4 | 1283.6 to 1297.7 | -44.5 | 18,333 |
| 17 | 1283.6 | gemini-3-flash | 1277.9 to 1289.2 | -51.5 | 33,678 | |
| 18 | 1282.7 | Moonshot AI | kimi-k2.6 | 1275.4 to 1290.0 | -52.4 | 15,215 |
| 19 | 1282.4 | Anthropic | claude-sonnet-4-6 | 1275.9 to 1289.0 | -52.7 | 23,147 |
| 20 | 1275.5 | ByteDance | dola-seed-2.0-pro | 1267.8 to 1283.2 | -59.6 | 10,345 |
| 21 | 1274.5 | Anthropic | claude-sonnet-5-high | 1263.8 to 1285.2 | -60.6 | 4,103 |
| 22 | 1272.2 | Alibaba | qwen3.7-plus | 1262.6 to 1281.8 | -62.9 | 5,541 |
| 23 | 1269.9 | gemma-4-31b | 1263.2 to 1276.6 | -65.2 | 31,087 | |
| 24 | 1267.4 | OpenAI | gpt-5.2-chat-latest-20260210 | 1260.5 to 1274.4 | -67.7 | 15,536 |
| 25 | 1266.7 | Alibaba | qwen3.5-397b-a17b | 1260.4 to 1273.0 | -68.4 | 22,096 |
Frontier race
Category Kings: Vision arena frontier by organization
Anthropic holds the vision arena frontier at 1335 as of July 12, 2026.
Method: Frontier is the maximum resolved LMArena rating per organization on each publish date. Organizations limited to the union of the seven tracked homepage providers and organizations in the latest board top 25 by officialRank. Quarantined dates are excluded, and this category series keeps its own method epoch and is never joined to the legacy archive.
| Organization | Latest model | Latest LMArena rating | First tracked date |
|---|---|---|---|
| Anthropic | anthropic/claude-fable-5 | 1335.1 | June 29, 2024 |
| google/gemini-3.5-flash-medium | 1308.8 | June 29, 2024 | |
| Meta | meta/muse-spark | 1307.3 | May 11, 2025 |
| OpenAI | openai/gpt-5.4-high | 1298.8 | June 29, 2024 |
| Moonshot AI | moonshot/kimi-k2.6 | 1282.7 | January 29, 2026 |
| ByteDance | bytedance/dola-seed-2.0-pro | 1275.5 | February 16, 2026 |
| Alibaba | alibaba/qwen3.7-plus | 1272.2 | September 30, 2025 |
| xAI | xai/grok-4.20-beta-0309-reasoning | 1259.3 | July 17, 2025 |
| Mistral AI | mistral/mistral-large-3 | 1225.4 | June 24, 2025 |
Definitions
What these terms mean.
- Vision arena
- The LMArena vision arena, where models answer prompts that include image input. Model Gauntlet mirrors this named board and adds no separate vision test.
- Frontier by organization
- The highest resolved score for each organization on each publish date. It tracks how far each lab has pushed this category, not an average of all its models.
- Confidence interval
- The reported 95% range LMArena publishes with each rating. Overlapping intervals mean this snapshot does not separate those ranks decisively.
Follow the evidence
Compare this board against the overall LMArena rankings, shortlist candidates with the Multimodal shortlist guide,read the Anthropic lab file, and see how rankings are selected. Model profiles on this board: Claude Fable 5, Claude Opus 4.7, Muse Spark.
Vision arena questions
Which model leads the Vision arena right now?
According to the LMArena vision board published July 12, 2026, Anthropic leads with claude-fable-5 at 1335.1, across 4,383 battles. That is 16.7 points ahead of the next board entry inside this one snapshot.
Is this a Model Gauntlet vision ranking?
No. This board mirrors one named source, the LMArena vision arena, published July 12, 2026. Model Gauntlet publishes no composite or consensus score and runs no evaluation of its own.
How many models and organizations are in the vision board?
The reviewed artifact covers 143 models across 29 organizations, with 113 models resolved to exact identities and 30 held for review. The history spans 140 snapshots from June 29, 2024 to July 12, 2026.
Why do some scores have overlapping confidence intervals?
LMArena publishes a 95% confidence interval with each rating. Where two intervals overlap, this single snapshot does not separate those ranks decisively, so treat small vision gaps with caution.