← All LMArena Text Arena snapshots

LMArena Text Arena snapshot: 20 Oct 2026

Permanent, dated copy. Two views of the same date, each from its own source. Live leaderboard →

Download CSV Both views, with a “view” column. Every row carries the source date, check dates and source link.

Text Arena Overall: default view, as ChatJudge copies it

On the LMArena Text Arena board dated 20 Oct 2026 (default view), Gemini 4 Argon (High) (Google) ranked #1 with an Arena score of 1524 (±9), marked Preliminary on the page.

What changed: This is the earliest snapshot of the default view in ChatJudge’s archive, so there is nothing earlier in this view to compare it with.

Arena and category
Text Arena, Overall
Style control
Not stated by source. This is the page’s default view (what ChatJudge copies), not the “style control off” view
Source date
20 Oct 2026, as printed on the LMArena leaderboard page
Checked by ChatJudge
22 Oct 2026 to 23 Oct 2026
Rows
29 of 417 models listed. Top 25 overall, plus the best-ranked entry for OpenAI, xAI, DeepSeek and Z.ai (ranks 26–39).
Source
LMArena Text leaderboard ↗
Text Arena Overall, default view, 20 Oct 2026
RankModelProviderArena scoreOutput price ($ per 1M tokens)±VotesInput price ($/1M)
1Gemini 4 Argon (High) PreliminaryGoogle1524$10±94,900$2
2Muse Wonder (Max)Meta1510$20±95,099$4
3Claude Opus 5.5 (High)Anthropic1509$20±79,496$4
4Claude Fable 5 (High)Anthropic1504$50±443,399$10
5Claude Opus 4.6 (High)Anthropic1503$25±382,619$5
6Claude Fable 5.1 (Max)Anthropic1503$50±615,623$10
7Claude Opus 4.7 (High)Anthropic1501$25±465,131$5
8Claude Opus 4.6Anthropic1497$25±387,372$5
9Gemini 3.8 Flash (High) PreliminaryGoogle1496$3.75±541,423$0.75
10Muse Spark 1.3 (Max)Meta1495$4.25±519,323$1.25
11Claude Opus 4.7Anthropic1494$25±466,185$5
12Muse Spark 1.2 (xHigh)Meta1492$4.25±79,578$1.25
13Claude Opus 5 (Max)Anthropic1491$25±533,683$5
14Muse Spark 1.1Meta1490$4.25±444,815$1.25
15Claude Opus 5 (High)Anthropic1490$25±469,307$5
16Muse SparkMeta1489–±614,153N/A
17Gemini 3.7 Flash (High) PreliminaryGoogle1489$3.75±529,177$0.75
18Kimi K3 (Max)Moonshot1488–±534,325N/A
19Gemini 3.1 Pro PreviewGoogle1487$6.00±3129,523$1
20Gemini 3 ProGoogle1485$12±442,059$2
21GPT-5.6 Sol (xHigh)OpenAI1484$20±441,803$4
22Gemini 3.6 Flash (High)Google1483$3.75±436,079$0.75
23Qwen3.8-MaxAlibaba1483$6.00±529,418$2
24GPT-5.5 (High)OpenAI1481$30±469,340$5
25Claude Opus 4.8 (High)Anthropic1481$25±464,063$5
26GLM-5.3 (Max)Z.ai1481$4.40±524,035$1.4
28GPT-6.1 Sol (Max)OpenAI1479$10±87,453$2
36DeepSeek-V4.1-Flash-20260910 (Max)DeepSeek1475$1.20±613,444$0.3
39Grok 4.20 Beta 1xAI1475–±527,918N/A

Text Arena Overall: style control off, from Arena’s published dataset

On the Text Arena board Arena published on 20 Oct 2026 (style control off), Gemini 4 Argon (High) (Google) ranked #1 with an Arena rating of 1532.3 (lower 1523.5, upper 1541.1).

What changed: This is the earliest snapshot of the style-control-off view in ChatJudge’s archive, so there is nothing earlier in this view to compare it with.

Arena and category
Text Arena, Overall
Style control
Off: Arena’s dataset config “text” (on arena.ai this is the “overall-no-style-control” view)
Source date
20 Oct 2026, the leaderboard_publish_date in Arena’s dataset
Read by ChatJudge
15 Nov 2026 (dataset file SHA-256 8f6282c5efc4…)
Rows
31 of 417 models listed. Top 25 overall, plus the best-ranked entry for each major provider outside the top 25. Ratings rounded to 0.1; lower and upper bounds as published in the dataset.
Source
LMArena leaderboard dataset (Hugging Face) ↗
Same view on arena.ai
Text Arena, no style control ↗
Text Arena Overall, style control off, 20 Oct 2026
RankModelOrganisation (as published)Arena ratingLowerUpperVotes
1Gemini 4 Argon (High)google1532.31523.51541.14,900
2Claude Opus 5.5 (High)anthropic1516.21509.21523.29,496
3Claude Fable 5.1 (Max)anthropic1511.91506.01517.915,623
4Claude Opus 5 (Max)anthropic1508.61503.91513.333,683
5Claude Opus 5 (High)anthropic1502.81499.11506.469,307
6Muse Wonder (Max)meta1501.71493.01510.45,099
7Claude Opus 4.6 (High)anthropic1501.21497.91504.582,619
8Gemini 3.8 Flash (High)google1498.91494.41503.441,423
9Claude Opus 4.6anthropic1497.01493.81500.287,372
10Claude Fable 5 (High)anthropic1491.31487.01495.543,399
11MiMo-V2.6-Proxiaomi1491.11481.71500.44,065
12Muse Spark 1.3 (Max)meta1490.51485.01495.919,323
13Claude Opus 4.7 (High)anthropic1489.71485.91493.465,131
14Gemini 3.7 Flash (High)google1488.51483.81493.329,177
15Qwen3.8-Maxalibaba1483.21478.31488.129,418
16Claude Opus 4.7anthropic1483.01479.31486.866,185
17Muse Spark 1.2 (xHigh)meta1482.41475.61489.19,578
18Gemini 3.5 Flash (High)google1482.01478.11485.848,479
19Gemini 3.1 Pro Previewgoogle1480.11477.11483.0129,523
20Gemini 3.6 Flash (High)google1479.51475.11483.936,079
21Gemini 3 Progoogle1479.51475.71483.242,059
22Muse Spark 1.1meta1478.31474.21482.544,815
23Gemini 3.5 Flash (Medium)google1477.61473.51481.847,228
24Kimi K3 (Max)moonshot1475.71471.21480.234,325
25Qwen3.7-Max-Previewalibaba1474.91465.21484.63,942
27GLM-5.3 (Max)zai1473.71468.61478.824,035
30GPT-5.5 (High)openai1471.41467.71475.069,340
38DeepSeek-V4.1-Flash-20260910 (Max)deepseek1463.61457.51469.713,444
51Grok 4.20 Beta (Reasoning, 0309)xai1450.81447.11454.566,403
91MiniMax-M3minimax1432.61428.91436.465,792
93Mistral Large 4mistral1430.81424.11437.49,308

Text Arena is run and published by LMArena (lmarena.ai, arena.ai), including the dated leaderboard dataset on Hugging Face. ChatJudge copies the published numbers and does not run the benchmark. Scores from different views or benchmarks aren’t directly comparable.