LMArena Document arena · published July 12, 2026
Anthropic leads the LMArena document arena.
As of July 12, 2026, Anthropic leads the LMArena document arena with claude-opus-4-6-thinking at 1508.1 across 24,188 battles, according to LMArena.
Current through July 12, 2026. 14 published snapshots. 34 of 34 models resolved (0 held for exact-match review).
| Rank | LMArena rating | Organization | Model | Reported interval | Gap to leader | Battles |
|---|---|---|---|---|---|---|
| 01 | 1508.1 | Anthropic | claude-opus-4-6-thinking | 1501.1 to 1515.0 | Leader | 24,188 |
| 02 | 1507.1 | Anthropic | Claude Fable 5 | 1495.5 to 1518.7 | -1.0 | 2,644 |
| 03 | 1506.7 | Anthropic | claude-opus-4-6 | 1500.5 to 1512.9 | -1.4 | 36,602 |
| 04 | 1504.5 | Anthropic | claude-opus-4-7-thinking | 1497.7 to 1511.2 | -3.6 | 18,023 |
| 05 | 1501.2 | Anthropic | Claude Opus 4.7 | 1494.7 to 1507.7 | -6.9 | 18,185 |
| 06 | 1488.2 | OpenAI | gpt-5.5-high | 1481.4 to 1494.9 | -19.9 | 15,996 |
| 07 | 1485.9 | Anthropic | claude-sonnet-4-6 | 1479.9 to 1491.8 | -22.2 | 53,846 |
| 08 | 1480.6 | OpenAI | gpt-5.5 | 1474.2 to 1487.1 | -27.5 | 16,410 |
| 09 | 1474.7 | Anthropic | claude-opus-4-8-thinking | 1466.9 to 1482.6 | -33.4 | 7,546 |
| 10 | 1472.2 | OpenAI | gpt-5.4 | 1466.0 to 1478.5 | -35.9 | 28,910 |
| 11 | 1468.9 | Anthropic | claude-opus-4-8 | 1460.7 to 1477.2 | -39.2 | 7,298 |
| 12 | 1468.8 | Anthropic | claude-sonnet-5-high | 1457.4 to 1480.3 | -39.3 | 2,809 |
| 13 | 1462.2 | gemini-3.5-flash-medium | 1450.6 to 1473.9 | -45.9 | 2,697 | |
| 14 | 1461.4 | Anthropic | claude-opus-4-5-20251101 | 1451.1 to 1471.7 | -46.7 | 7,985 |
| 15 | 1449.2 | Moonshot AI | kimi-k2.6 | 1441.6 to 1456.8 | -58.9 | 11,094 |
| 16 | 1445.8 | Anthropic | claude-sonnet-4-5-20250929 | 1439.4 to 1452.2 | -62.3 | 27,977 |
| 17 | 1444.9 | Meta | Muse Spark | 1426.8 to 1463.0 | -63.2 | 1,086 |
| 18 | 1444.2 | Alibaba | qwen3.7-plus | 1433.0 to 1455.3 | -63.9 | 2,700 |
| 19 | 1441.4 | gemini-3.1-pro-preview | 1436.0 to 1446.8 | -66.7 | 44,046 | |
| 20 | 1435.4 | MiniMax | minimax-m3 | 1427.1 to 1443.6 | -72.7 | 6,328 |
| 21 | 1433.6 | Gemini 3 Pro | 1424.8 to 1442.4 | -74.5 | 10,748 | |
| 22 | 1430.6 | Moonshot AI | kimi-k2.5-thinking | 1423.8 to 1437.3 | -77.5 | 19,342 |
| 23 | 1424.2 | gemma-4-31b | 1416.1 to 1432.4 | -83.9 | 10,132 | |
| 24 | 1421.7 | gemini-2.5-pro | 1415.5 to 1427.9 | -86.4 | 25,053 | |
| 25 | 1420.9 | Anthropic | claude-haiku-4-5-20251001 | 1414.8 to 1427.0 | -87.2 | 30,187 |
Frontier race
Category Kings: Document arena frontier by organization
Anthropic holds the document arena frontier at 1508 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-opus-4-6-thinking | 1508.1 | March 3, 2026 |
| OpenAI | openai/gpt-5.5-high | 1488.2 | March 3, 2026 |
| google/gemini-3.5-flash-medium | 1462.2 | March 3, 2026 | |
| Moonshot AI | moonshot/kimi-k2.6 | 1449.2 | April 14, 2026 |
| Meta | meta/muse-spark | 1444.9 | April 19, 2026 |
| Alibaba | alibaba/qwen3.7-plus | 1444.2 | July 2, 2026 |
| MiniMax | minimax/minimax-m3 | 1435.4 | June 3, 2026 |
| xAI | xai/grok-4.20-beta-0309-reasoning | 1413.1 | April 14, 2026 |
Definitions
What these terms mean.
- Document arena
- The LMArena document arena, where models answer prompts grounded in document input. Model Gauntlet mirrors this named board without adding its own document 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.
Document arena questions
Which model leads the Document arena right now?
According to the LMArena document board published July 12, 2026, Anthropic leads with claude-opus-4-6-thinking at 1508.1, across 24,188 battles. That is 1.0 points ahead of the next board entry inside this one snapshot.
Is this a Model Gauntlet document ranking?
No. This board mirrors one named source, the LMArena document 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 document board?
The reviewed artifact covers 34 models across 9 organizations, with 34 models resolved to exact identities and 0 held for review. The history spans 14 snapshots from March 3, 2026 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 document gaps with caution.