BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases8

LMArena Multi-turn leaderboard

As of 5 Nov 2026, the top score on LMArena Multi-turn is 1559 by Gemini 4 Carbon (Google), ahead of Gemini 4 Argon at 1539, across 374 model configurations with a published result (Arena (LMArena), result dated 28 Oct 2026). Pairwise votes on multi-turn conversations, rewarding models that track context. LMArena Multi-turn is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 5 Nov 2026)

  • Gemini 4 Carbon holds the top Arena Multi-turn score at 1559 (Arena (LMArena), 28 Oct 2026).
  • Gemini 4 Argon is second on Arena Multi-turn at 1539 (Arena (LMArena), 28 Oct 2026).
  • Arena Multi-turn: Text arena rating on conversations with more than one turn (Arena (LMArena)).
  • 374 model configurations have a published Arena Multi-turn result on BenchLeader (BenchLeader, 5 Nov 2026).
  • Arena Multi-turn does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Text arena rating on conversations with more than one turn.

Published by
Arena (LMArena)
Category
Human preference
Index weight
Reference only
Models
374
Data as of
5 Nov 2026

CC BY 4.0 — lmarena-ai/leaderboard-dataset on Hugging Face.

What the test looks like

Pairwise votes on multi-turn conversations, rewarding models that track context.

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 overall board.

Top 15
  1. 1Gemini 4 Carbon (high)1559
  2. 2Gemini 4 Argon (high)1539
  3. 3Claude Opus 4.6 (high)1517
  4. 4Muse Wonder (max)1517
  5. 5Claude Fable 5 (high)1516
  6. 6Claude Opus 4.7 (high)1516
  7. 7Claude Opus 4.71514
  8. 8Claude Opus 4.61510
  9. 9Muse Spark 1.2 (xhigh)1508
  10. 10Gemini 3.8 Flash (high)1505
  11. 11Kimi K3 (max)1502
  12. 12Claude Opus 5.5 (high)1499
  13. 13Claude Opus 4.8 (high)1497
  14. 14Gemini 3.1 Pro1496
  15. 15Gemini 3 Pro1496
374 of 374
#
1NewGemini 4 Carbonhigh155969.12026-10-28
2Gemini 4 Argonhigh153965.92026-10-28
3Claude Opus 4.6high151762.62026-10-28
4Muse Wondermax151762.22026-10-28
5Claude Fable 5high151664.32026-10-28
6Claude Opus 4.7high151660.92026-10-28
7Claude Opus 4.7151463.12026-10-28
8Claude Opus 4.6151062.32026-10-28
9Muse Spark 1.2xhigh150861.72026-10-28
10Gemini 3.8 Flashhigh150562.02026-10-28
11Kimi K3max150262.12026-10-28
12Claude Opus 5.5high149965.82026-10-28
13Claude Opus 4.8high149760.62026-10-28
14Gemini 3.1 Pro149660.52026-10-28
15Gemini 3 Pro149660.22026-10-28
16Muse Spark 1.1149662.22026-10-28
17Claude Opus 4.8149460.02026-10-28
18GPT-5.2149457.92026-10-28
19Qwen3.8 Maxmax149458.82026-10-28
20Muse Spark 1.3max149361.92026-10-28
21Muse Spark149364.22026-10-28
22GPT-5.4high149358.82026-10-28
23GPT-5.6 Solxhigh149064.22026-10-28
24Claude Fable 5.1max148967.22026-10-28
25Gemini 3.7 Flashhigh148862.22026-10-28
26Gemini 3.6 Flashhigh148861.32026-10-28
27GPT-5.5high148763.72026-10-28
28Claude Opus 5high148764.32026-10-28
29Claude Opus 5max148665.42026-10-28
30Claude Opus 4.5high148559.12026-10-28
31Claude Opus 4.5148457.92026-10-28
32GPT-6 Astramax148467.82026-10-28
33GPT-5.5148461.72026-10-28
34Grok 4.20 11483–2026-10-28
35GPT-5.5 Instant148357.02026-10-28
36Gemini 3 Flash148257.02026-10-28
37Qwen3.7max148260.32026-10-28
38Claude Sonnet 4.6148158.32026-10-28
39Gemini 3.5 Flashhigh148062.62026-10-28
40GPT-5.4148058.02026-10-28
41MiMo-V2.5-Pro148054.92026-10-28
42GPT-6.1 Solmax147967.22026-10-28
43Grok 4.20thinking147958.42026-10-28
44GLM 5.1147855.82026-10-28
45GLM 5.3max147859.02026-10-28
46Gemini 3.5 Flashmedium147861.92026-10-28
47Qwen3.5 Maxmax1477–2026-10-28
48Claude Sonnet 4.5147753.72026-10-28
49Claude Sonnet 5.5xhigh147763.42026-10-28
50GLM 5.2max147660.22026-10-28
51GLM 5.3 Flash147656.82026-10-28
52DeepSeek V4 Pro147553.22026-10-28
53GPT-5.6 Terraxhigh147460.32026-10-28
54Grok 4.20 Multi-Agent147458.42026-10-28
55Gemini 3 Flashminimal147453.92026-10-28
56Grok 4.5147359.82026-10-28
57Claude Sonnet 5high147259.32026-10-28
58Claude Sonnet 4.5high147257.92026-10-28
59Claude Opus 4.1thinking147157.92026-10-28
60GLM-5147154.82026-10-28

More human preference benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on Arena Multi-turn

Cite as: BenchLeader, “Arena Multi-turn leaderboard”, data as of 5 Nov 2026, https://www.benchleader.com/benchmarks/lmarena_text_multiturn.

Arena Multi-turn: questions

What does Arena Multi-turn measure?
Pairwise votes on multi-turn conversations, rewarding models that track context. Scores are reported in Elo-style ratings from pairwise comparisons; higher is better.
What is the top score on LMArena Multi-turn?
As of 5 Nov 2026, the highest score on Arena Multi-turn is 1559 by Gemini 4 Carbon (result dated 28 Oct 2026), ahead of Gemini 4 Argon at 1539. Scores come from Arena (LMArena) and are re-read daily.
Which AI model leads Arena Multi-turn?
Gemini 4 Carbon leads Arena Multi-turn with 1559 as of 5 Nov 2026 (result dated 28 Oct 2026), ahead of Gemini 4 Argon at 1539.
How many models have Arena Multi-turn results?
374 model configurations have a Arena Multi-turn result on BenchLeader, all taken from Arena (LMArena).
Who runs Arena Multi-turn and how often is it updated?
Arena Multi-turn is published by Arena (LMArena). BenchLeader re-reads the published results once a day and records the date each result was published.
Does Arena Multi-turn count toward the BenchLeader Index?
No. Arena Multi-turn is shown for reference but left out of the composite index.