LMArena Document leaderboard
As of 6 Nov 2026, the top score on LMArena Document is 1499 by Claude Fable 5 (Anthropic), ahead of Claude Opus 4.6 at 1495, across 44 model configurations with a published result (Arena (LMArena), result dated 13 Sep 2026). Document Arena compares models answering questions over an uploaded file. LMArena Document is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- Claude Fable 5 holds the top Arena Document score at 1499 (Arena (LMArena), 13 Sep 2026).
- Claude Opus 4.6 is second on Arena Document at 1495 (Arena (LMArena), 13 Sep 2026).
- Arena Document: Pairwise votes on answers about uploaded documents (Arena (LMArena)).
- 44 model configurations have a published Arena Document result on BenchLeader (BenchLeader, 6 Nov 2026).
- Arena Document does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Pairwise votes on answers about uploaded documents.
- Published by
- Arena (LMArena)
- Category
- Long context
- Index weight
- Reference only
- Models
- 44
- Data as of
- 6 Nov 2026
CC BY 4.0 — lmarena-ai/leaderboard-dataset on Hugging Face.
What the test looks like
Document Arena compares models answering questions over an uploaded file.
How it is scored
Bradley-Terry rating from pairwise votes, in Elo-like units, with style control.
What to keep in mind
Fewer votes than the text arena; document types are whatever users upload.
- 1Claude Fable 51499
- 2Claude Opus 4.6 (high)1495
- 3Claude Opus 4.71494
- 4Claude Opus 4.61494
- 5Claude Opus 4.7 (high)1493
- 6Claude Fable 5.1 (max)1493
- 7Claude Opus 5 (high)1490
- 8GPT-5.6 Sol (xhigh)1489
- 9Claude Opus 4.8 (high)1486
- 10GPT-5.51483
- 11GPT-5.5 (high)1480
- 12Muse Spark 1.11480
- 13Claude Sonnet 4.61478
- 14GPT-5.6 Terra (xhigh)1477
- 15Claude Sonnet 5 (high)1476
44 of 44
| # | ||||
|---|---|---|---|---|
| 1 |  Claude Fable 5 | 1499 | 66.5 | 2026-09-13 |
| 2 | Claude Opus 4.6high | 1495 | 62.6 | 2026-09-13 |
| 3 | Claude Opus 4.7 | 1494 | 63.1 | 2026-09-13 |
| 4 | Claude Opus 4.6 | 1494 | 62.3 | 2026-09-13 |
| 5 | Claude Opus 4.7high | 1493 | 60.9 | 2026-09-13 |
| 6 | Claude Fable 5.1max | 1493 | 67.2 | 2026-09-13 |
| 7 | Claude Opus 5high | 1490 | 64.3 | 2026-09-13 |
| 8 |  GPT-5.6 Solxhigh | 1489 | 64.2 | 2026-09-13 |
| 9 | Claude Opus 4.8high | 1486 | 60.6 | 2026-09-13 |
| 10 | GPT-5.5 | 1483 | 61.7 | 2026-09-13 |
| 11 | GPT-5.5high | 1480 | 63.7 | 2026-09-13 |
| 12 |  Muse Spark 1.1 | 1480 | 62.2 | 2026-09-13 |
| 13 | Claude Sonnet 4.6 | 1478 | 58.2 | 2026-09-13 |
| 14 | GPT-5.6 Terraxhigh | 1477 | 60.3 | 2026-09-13 |
| 15 | Claude Sonnet 5high | 1476 | 59.3 | 2026-09-13 |
| 16 | Claude Opus 4.8 | 1475 | 60.0 | 2026-09-13 |
| 17 | GPT-6 Astramax | 1473 | 67.8 | 2026-09-13 |
| 18 | GPT-5.4 | 1470 | 58.0 | 2026-09-13 |
| 19 | Muse Spark 1.3max | 1468 | 62.0 | 2026-09-13 |
| 20 | Muse Spark | 1467 | 64.1 | 2026-09-13 |
| 21 | Claude Opus 4.5 | 1465 | 57.9 | 2026-09-13 |
| 22 |  Grok 4.5 | 1464 | 59.8 | 2026-09-13 |
| 23 | GPT-5.6 Lunaxhigh | 1462 | 58.5 | 2026-09-13 |
| 24 | Grok 4.6high | 1461 | 60.5 | 2026-09-13 |
| 25 |  Gemini 3.5 Flashmedium | 1461 | 61.9 | 2026-09-13 |
| 26 | GPT-5.5 Instant | 1460 | 57.0 | 2026-09-13 |
| 27 | Gemini 3.1 Pro | 1459 | 60.5 | 2026-09-13 |
| 28 | Gemini 3.5 Flashhigh | 1453 | 62.5 | 2026-09-13 |
| 29 | Gemini 3.6 Flashhigh | 1453 | 61.2 | 2026-09-13 |
| 30 |  Qwen3.7 Plus | 1453 | 57.7 | 2026-09-13 |
| 31 | Gemini 3 Pro | 1451 | 60.2 | 2026-09-13 |
| 32 |  Kimi K2.6open ↗ | 1450 | 55.2 | 2026-09-13 |
| 33 | Claude Sonnet 4.5 | 1449 | 53.7 | 2026-09-13 |
| 34 | Gemma 4 31Bopen ↗ | 1445 | 47.8 | 2026-09-13 |
| 35 | Grok 4.20thinking | 1439 | 58.4 | 2026-09-13 |
| 36 | Gemini 3 Flash | 1439 | 56.9 | 2026-09-13 |
| 37 | Kimi K2.5thinkingopen ↗ | 1436 | 58.7 | 2026-09-13 |
| 38 |  MiniMax M3open ↗ | 1433 | 56.4 | 2026-09-13 |
| 39 | Gemini 2.5 Pro | 1431 | 54.5 | 2026-09-13 |
| 40 | GPT-5.2 | 1425 | 57.9 | 2026-09-13 |
| 41 | GPT-5.2high | 1423 | 56.0 | 2026-09-13 |
| 42 | Claude Haiku 4.5 | 1419 | 47.6 | 2026-09-13 |
| 43 | GPT-5.1 | 1416 | 56.8 | 2026-09-13 |
| 44 |  GLM 5V Turbo | 1400 | 55.8 | 2026-09-13 |
More long context benchmarks · All LLM benchmarks
Related benchmarks
Top head-to-heads on Arena Document
Cite as: BenchLeader, “Arena Document leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/lmarena_document.
Arena Document: questions
- What does Arena Document measure?
- Document Arena compares models answering questions over an uploaded file. Scores are reported in Elo-style ratings from pairwise comparisons; higher is better.
- What is the top score on LMArena Document?
- As of 6 Nov 2026, the highest score on Arena Document is 1499 by Claude Fable 5 (result dated 13 Sep 2026), ahead of Claude Opus 4.6 at 1495. Scores come from Arena (LMArena) and are re-read daily.
- Which AI model leads Arena Document?
- Claude Fable 5 leads Arena Document with 1499 as of 6 Nov 2026 (result dated 13 Sep 2026), ahead of Claude Opus 4.6 at 1495.
- How many models have Arena Document results?
- 44 model configurations have a Arena Document result on BenchLeader, all taken from Arena (LMArena).
- Who runs Arena Document and how often is it updated?
- Arena Document is published by Arena (LMArena). BenchLeader re-reads the published results once a day and records the date each result was published.
- Does Arena Document count toward the BenchLeader Index?
- No. Arena Document is shown for reference but left out of the composite index.