Skip to main content
AI Choice Engine
← Benchmark sources

arena.ai · benchmark source

Arena

Live human-preference and agent boards from Arena (LMArena). Scores are board-specific — Text Arena Score, WebDev Arena Score, and Agent net improvement are not interchangeable. The Arena product home at arena.ai/ and the long-running lmarena.ai leaderboards are the same operator (verified 2026-08-29 via arena.ai root and the LMArena research pages; see https://lmarena.ai/ for the historic domain).

Open arena.ai3 boards on this publisher

General chat

Text Arena · Overall

Arena's overall text-to-text preference leaderboard across math, coding, creative writing, and other open-ended tasks, scored with Arena Score (Bradley-Terry / Elo-style).

Why it matters

It is the broadest live preference signal for general chat quality when you need an orienting shortlist before task-specific benchmarks.

Use it when

You are shortlisting models for general assistant or chat work and want a live preference signal with honest coverage limits.

The limitation

Votes reflect crowd taste on open-ended prompts, not factual accuracy, agent tool use, or enterprise safety. Arena Score is board-specific and not comparable to WebDev or Agent net improvement.

Leaderboard snapshot

Text Arena · Overall results in a readable view

Use the tabs for the primary score and other published fields from this board. The source link remains authoritative for the live table.

Arena Text capture (410 ranked rows, board dated 2026-09-30). Top: gemini-4-argon-high (1525), claude-fable-5-high (1505), claude-opus-4-6-high (1505) — the top cluster sits inside overlapping confidence intervals.

Leader among verified rows

gemini-4-argon-high

1,525

Rows shown

119

Arena Score

Source / snapshot date

2026-10-01

25 days old · not a live API feed

Score profile

Arena Score by published row

Higher is better in this view

gemini-4-argon-high
1,525
claude-fable-5-high
1,505
claude-opus-4-6-high
1,505
claude-opus-5.5-high
1,504
claude-opus-4-7-high
1,502
claude-fable-5.1-max
1,501
claude-opus-4-6
1,497
muse-spark-1.3-max
1,495
claude-opus-4-7
1,494
gemini-3.8-flash-high
1,494
muse-spark-1.2 (xHigh)
1,494
muse-spark-1.1
1,492
claude-opus-5-high
1,491
claude-opus-5-max
1,489
muse-spark
1,489
gemini-3.7-flash-high
1,488
kimi-k3-max
1,488
gemini-3.1-pro-preview
1,487
gemini-3-pro
1,485
gpt-5.6-sol-xhigh
1,484
gemini-3.6-flash-high
1,483
claude-opus-4-8-high
1,481
gpt-5.5-high
1,481
qwen3.8-max
1,481
mimo-v2.6-pro
1,480
glm-5.3-max
1,479
gemini-3.5-flash-high
1,477
gpt-5.5
1,477
gemini-3.5-flash-medium
1,476
gpt-5.2-chat-latest-20260210
1,476
gpt-6-astra-max
1,476
claude-opus-4-8
1,475
glm-5.2-max
1,475
gpt-5.4-high
1,475
grok-4.20-beta1
1,475
qwen3.7-max-preview
1,475
claude-opus-4-5-20251101-high-32k
1,474
glm-5.3-flash
1,474
deepseek-v4.1-flash-max
1,473
gemini-3-flash
1,473
gpt-5.5-instant
1,473
claude-sonnet-4-6
1,472
grok-4.20-beta-0309-reasoning
1,472
grok-4.20-multi-agent-beta-0309
1,471
claude-opus-4-5-20251101
1,470
ernie-5.1
1,468
mimo-v2.5-pro
1,467
grok-4.5
1,466
glm-5.1
1,465
gpt-5.4
1,465
gpt-5.6-terra-xhigh
1,465
grok-4.1-thinking
1,465
qwen3.5-max-preview
1,465
deepseek-v4-pro-high-20260813
1,464
claude-sonnet-5-high
1,462
kimi-k2.6
1,461
qwen3.6-max-preview
1,460
grok-4.1
1,459
deepseek-v4-pro
1,458
gemini-3-flash (thinking-minimal)
1,458
glm-5
1,458
claude-sonnet-4-5-20250929-high-32k
1,457
dola-seed-2.0-pro
1,457
hy3
1,457
gpt-5.1-high
1,456
gpt-6-sol-max
1,456
qwen3.7-plus
1,456
claude-sonnet-4-5-20250929
1,455
deepseek-v4-pro-high-preview
1,455
gemini-3.5-flash-lite
1,454
mimo-v2.6-flash
1,454
gemma-4-31b
1,453
gpt-5.6-luna-xhigh
1,453
grok-4.6-high
1,453
claude-opus-4-1-20250805-thinking-16k
1,450
gpt-5.3-chat-latest
1,450
kimi-k2.5-thinking
1,450
ernie-5.0-preview-1203
1,449
claude-opus-4-1-20250805
1,448
mimo-v2-pro
1,448
ernie-5.0-0110
1,447
gpt-5.4-mini-high
1,447
gemini-2.5-pro
1,446
gpt-4.5-preview-2025-02-27
1,445
gpt-6-luna-max
1,444
chatgpt-4o-latest-20250326
1,443
qwen3.6-plus
1,443
glm-4.7
1,442
grok-4.7-xhigh
1,442
inkling
1,442
qwen3.5-397b-a17b
1,442
grok-4.3
1,441
minimax-m3
1,440
qwen3.8-27b
1,439
deepseek-v4-flash-high-preview
1,438
gpt-5.1
1,438
gemma-4-26b-a4b
1,437
gpt-5.2-high
1,437
longcat-flash-chat-2602-exp
1,437
deepseek-v4-flash
1,436
gpt-5.2
1,436
mimo-v2.5
1,434
qwen3-max-preview
1,434
gemini-3.1-flash-lite-preview
1,433
mistral-medium-3.5
1,427
qwen3-max-2025-09-23
1,423
deepseek-r1-0528
1,422
qwen3-235b-a22b-instruct-2507
1,422
deepseek-v3.1
1,417
deepseek-v3.1-thinking
1,416
minimax-m2.7
1,415
claude-haiku-4-5-20251001
1,414
mistral-large-3
1,414
Inkling Small
1,405
gpt-5.4-nano-high
1,401
qwen3-235b-a22b-thinking-2507
1,400
deepseek-r1
1,398
minimax-m2.5
1,391
qwen3-235b-a22b
1,375

Full Text Arena overall table captured 2026-10-01 from arena.ai/leaderboard/text (410 models, 8,602,501 votes; board dated 2026-09-30). Confidence intervals overlap heavily below the top cluster. Arena Score is board-specific and not comparable to WebDev, Agent net improvement, or SWE-bench percentages.

Open Text Arena leaderboard

Showing the top 100 of 410 entries plus every model with a page in our catalog. 291 lower-ranked entries are on the full source table.

Evidence in this catalog

Where Text Arena · Overall fits

Closest verified examples we currently carry. A missing score is not a zero.

Model examples

  • Published source observation captured 2026-10-01. Restricted preview; a leaderboard run does not establish general API availability.

  • Published source observation captured 2026-10-01. Source-specific result; no independent rerun or universal capability verdict.

  • Published source observation captured 2026-10-01. Source-specific result; no independent rerun or universal capability verdict.

Harness examples

No directly comparable harness score is verified here yet.

Open this board on Text Arena leaderboard