Benchmark

LMArena (text)

LMArena · source page ↗ · last checked Jul 5, 2026, 6:02 PM

Current leaderboard

captured Jul 5, 2026, 6:02 PM

Model	Arena Elo (text, overall)
claude-opus-4-6-thinking	1501
claude-opus-4-6	1498
claude-fable-5	1495
claude-opus-4-7-thinking	1489
gemini-3.5-flash	1482
claude-opus-4-7	1482
gemini-3.1-pro-preview	1480
gemini-3-pro	1479
qwen3.7-max-preview	1475
muse-spark	1473
qwen3.5-max-preview	1470
gpt-5.4-high	1470
gpt-5.5-high	1469
glm-5.1	1467
ernie-5.1	1467
gemini-3-flash	1466
glm-5.2 (max)	1463
gpt-5.5	1463
claude-opus-4-8-thinking	1463
mimo-v2.5-pro	1462
qwen3.7-plus	1462
gemini-2.5-pro	1457
claude-sonnet-4-6	1457
kimi-k2.6	1455
claude-opus-4-8	1454

Change history

Benchmark Jul 3, 2026, 6:02 AM

LMArena (text) scores changed

Arena Elo (text, overall) updated.

View raw diff +21 −21

- claude-fable-5 1494
- claude-opus-4-6-thinking 1500
- claude-opus-4-8 1456
- deepseek-v4-pro 1449
- deepseek-v4-pro-thinking 1446
- gemini-3.5-flash 1480
- gemma-4-26b-a4b 1435
- glm-5 1446
- glm-5.1 1468
- glm-5.2 (max) 1465
- gpt-5.2 1411
- gpt-5.4 1454
- gpt-5.5-high 1468
- grok-4.1-thinking 1437
- grok-4.20-beta-0309-reasoning 1455
- mimo-v2-omni 1423
- mimo-v2.5 1427
- minimax-m3 1440
- muse-spark 1472
- nvidia-nemotron-3-ultra-550b-a55b-nvfp4 1439
- qwen3.7-plus 1463
+ claude-fable-5 1495
+ claude-opus-4-6-thinking 1501
+ claude-opus-4-8 1454
+ claude-sonnet-5-thinking 1445
+ deepseek-v4-pro 1450
+ deepseek-v4-pro-thinking 1447
+ gemini-3.5-flash 1482
+ gemma-4-26b-a4b 1434
+ glm-5 1445
+ glm-5.1 1467
+ glm-5.2 (max) 1463
+ gpt-5.4 1453
+ gpt-5.5-high 1469
+ grok-4.1-thinking 1438
+ grok-4.20-beta-0309-reasoning 1454
+ mimo-v2-omni 1422
+ mimo-v2.5 1426
+ minimax-m3 1439
+ muse-spark 1473
+ nvidia-nemotron-3-ultra-550b-a55b-nvfp4 1438
+ qwen3.7-plus 1462

Benchmark Jun 27, 2026, 6:01 AM

LMArena (text) scores changed

Arena Elo (text, overall) updated.

View raw diff +29 −29

- claude-opus-4-5-20251101 1449
- claude-opus-4-5-20251101-thinking-32k 1446
- claude-opus-4-6-thinking 1501
- claude-opus-4-7 1481
- claude-opus-4-8 1455
- claude-opus-4-8-thinking 1462
- claude-sonnet-4-6 1456
- deepseek-v4-flash 1430
- deepseek-v4-flash-thinking 1422
- deepseek-v4-pro 1448
- deepseek-v4-pro-thinking 1447
- gemini-3-flash (thinking-minimal) 1445
- gemma-4-26b-a4b 1434
- glm-5.1 1470
- glm-5.2 (max) 1467
- gpt-5.4-mini-high 1412
- grok-4-0709 1410
- grok-4.1 1437
- grok-4.20-beta-0309-reasoning 1453
- grok-4.20-multi-agent-beta-0309 1451
- kimi-k2.5-instant 1420
- kimi-k2.6 1454
- mimo-v2.5 1426
- minimax-m3 1442
- mistral-medium-3.5 1419
- nvidia-nemotron-3-ultra-550b-a55b-nvfp4 1435
- qwen3.5-397b-a17b 1440
- qwen3.6-max-preview 1447
- qwen3.6-plus 1438
+ claude-opus-4-5-20251101 1450
+ claude-opus-4-5-20251101-thinking-32k 1447
+ claude-opus-4-6-thinking 1500
+ claude-opus-4-7 1482
+ claude-opus-4-8 1456
+ claude-opus-4-8-thinking 1463
+ claude-sonnet-4-6 1457
+ deepseek-v4-flash 1431
+ deepseek-v4-flash-thinking 1423
+ deepseek-v4-pro 1449
+ deepseek-v4-pro-thinking 1446
+ gemini-3-flash (thinking-minimal) 1444
+ gemma-4-26b-a4b 1435
+ glm-5.1 1468
+ glm-5.2 (max) 1465
+ gpt-5.4-mini-high 1413
+ grok-4.1 1436
+ grok-4.20-beta-0309-reasoning 1455
+ grok-4.20-multi-agent-beta-0309 1450
+ kimi-k2.5-instant 1421
+ kimi-k2.6 1455
+ mimo-v2.5 1427
+ minimax-m3 1440
+ mistral-medium-3.5 1420
+ nvidia-nemotron-3-ultra-550b-a55b-nvfp4 1439
+ qwen3.5-397b-a17b 1439
+ qwen3.6-max-preview 1446
+ qwen3.6-plus 1437
+ qwen3.7-plus 1463