Model Markets
Get Free API KeyAPI

LMArena Text Arena

LMArena · text-2026-09-01-011508720696 · Blind pairwise human preferences from LMArena's text arena.

Model Ranking

Source
#1Claude Opus 5Anthropic1,504.96rating—16,839claude-opus-5-max; 95% CI [1499.16488972, 1510.76500615]; votes 16839; rank 1unknownOriginal sourcewinner eligibleThird-party benchmark
#2Claude Fable 5Anthropic1,494.40rating—26,977claude-fable-5; 95% CI [1489.42750707, 1499.38052801]; votes 26977; rank 5unknownOriginal sourcewinner eligibleThird-party benchmark
#3Gemini 3.7 FlashGoogle DeepMind1,491.10rating—5,685gemini-3.7-flash-high; 95% CI [1482.94370265, 1499.25525622]; votes 5685; rank 6unknownOriginal sourcewinner eligibleThird-party benchmark
#4Gemini 3.5 FlashGoogle DeepMind1,483.50rating—33,874gemini-3.5-flash-high; 95% CI [1479.17846588, 1487.82213514]; votes 33874; rank 9unknownOriginal sourcewinner eligibleThird-party benchmark
#5Gemini 3.1 ProGoogle DeepMind1,480.10rating—102,763gemini-3.1-pro-preview; 95% CI [1476.89767236, 1483.30025658]; votes 102763; rank 11unknownOriginal sourcewinner eligibleThird-party benchmark
#6Qwen3.8-MaxQwen1,479.56rating—13,136qwen3.8-max; 95% CI [1473.34740586, 1485.77213629]; votes 13136; rank 12unknownOriginal sourcewinner eligibleThird-party benchmark
#7Gemini 3.6 FlashGoogle DeepMind1,476.42rating—22,009gemini-3.6-flash-high; 95% CI [1471.33793256, 1481.50948001]; votes 22009; rank 15unknownOriginal sourcewinner eligibleThird-party benchmark
#8Kimi-K3Moonshot AI1,476.31rating—17,895kimi-k3-max; 95% CI [1470.86166728, 1481.76254962]; votes 17895; rank 16unknownOriginal sourcewinner eligibleThird-party benchmark
#9GLM-5.3Z.aiSelected model1,474.49rating—7,401glm-5.3-max; 95% CI [1467.12172899, 1481.86064533]; votes 7401; rank 18unknownOriginal sourcewinner eligibleThird-party benchmark
#10GLM-5.3-FlashZ.ai1,470.91rating—4,399glm-5.3-flash; 95% CI [1461.80526636, 1480.01283657]; votes 4399; rank 23unknownOriginal sourcewinner eligibleThird-party benchmark
#11ERNIE 5.1Baidu1,468.10rating—37,137ernie-5.1; 95% CI [1463.48233336, 1472.72453179]; votes 37137; rank 25unknownOriginal sourcewinner eligibleThird-party benchmark
#12GLM-5.2Z.ai1,466.80rating—33,758glm-5.2-max; 95% CI [1462.10739564, 1471.49788772]; votes 33758; rank 26unknownOriginal sourcewinner eligibleThird-party benchmark
#13Gemini 3 FlashGoogle DeepMind1,466.68rating—30,248gemini-3-flash; 95% CI [1462.28845512, 1471.07550892]; votes 30248; rank 28unknownOriginal sourcewinner eligibleThird-party benchmark
#14GLM-5.1Z.ai1,462.73rating—45,403glm-5.1; 95% CI [1458.76227206, 1466.68811970]; votes 45403; rank 30unknownOriginal sourcewinner eligibleThird-party benchmark
#15Claude Opus 4.8Anthropic1,461.14rating—48,467claude-opus-4-8-high; 95% CI [1456.88347873, 1465.40568534]; votes 48467; rank 31unknownOriginal sourcewinner eligibleThird-party benchmark
#16Gemini 2.5 ProGoogle DeepMind1,457.82rating—122,575gemini-2.5-pro; 95% CI [1455.34940240, 1460.28624287]; votes 122575; rank 33unknownOriginal sourcewinner eligibleThird-party benchmark
#17Kimi-K2.6Moonshot AI1,455.01rating—37,573kimi-k2.6; 95% CI [1450.48846324, 1459.52803920]; votes 37573; rank 34unknownOriginal sourcewinner eligibleThird-party benchmark
#18GPT-5.6 SolOpenAI1,454.69rating—23,153gpt-5.6-sol-xhigh; 95% CI [1449.58093424, 1459.80348090]; votes 23153; rank 35unknownOriginal sourcewinner eligibleThird-party benchmark
#19Qwen3.7-PlusQwen1,453.70rating—37,854qwen3.7-plus; 95% CI [1449.21174793, 1458.19774266]; votes 37854; rank 36unknownOriginal sourcewinner eligibleThird-party benchmark
#20DeepSeek-V4-ProDeepSeek1,451.00rating—54,243deepseek-v4-pro; 95% CI [1447.00829245, 1454.98663953]; votes 54243; rank 41unknownOriginal sourcewinner eligibleThird-party benchmark
#21Grok 4.20 Multi-AgentxAI1,450.22rating—60,845grok-4.20-multi-agent-beta-0309; 95% CI [1446.44654530, 1453.99996127]; votes 60845; rank 43unknownOriginal sourcewinner eligibleThird-party benchmark
#22GPT-5.6 TerraOpenAI1,446.72rating—23,942gpt-5.6-terra-xhigh; 95% CI [1441.68248185, 1451.74987487]; votes 23942; rank 47unknownOriginal sourcewinner eligibleThird-party benchmark
#23GLM-5Z.ai1,446.35rating—27,643glm-5; 95% CI [1442.03016476, 1450.67686961]; votes 27643; rank 48unknownOriginal sourcewinner eligibleThird-party benchmark
#24Kimi-K2.5Moonshot AI1,445.75rating—70,620kimi-k2.5-thinking; 95% CI [1442.41231565, 1449.09515875]; votes 70620; rank 50unknownOriginal sourcewinner eligibleThird-party benchmark
#25Grok 4.6xAI1,443.75rating—3,453grok-4.6-high; 95% CI [1433.62910542, 1453.86294509]; votes 3453; rank 55unknownOriginal sourcewinner eligibleThird-party benchmark
#26Claude Sonnet 5Anthropic1,442.95rating—31,096claude-sonnet-5-high; 95% CI [1438.22001632, 1447.67497803]; votes 31096; rank 56unknownOriginal sourcewinner eligibleThird-party benchmark
#27Hy3Tencent1,439.99rating—6,911hy3; 95% CI [1432.36272776, 1447.61573606]; votes 6911; rank 63unknownOriginal sourcewinner eligibleThird-party benchmark
#28Qwen3.5-397B-A17BQwen1,437.84rating—73,400qwen3.5-397b-a17b; 95% CI [1434.48605524, 1441.19621255]; votes 73400; rank 68unknownOriginal sourcewinner eligibleThird-party benchmark
#29Qwen3.8-27BQwen1,437.35rating—6,626qwen3.8-27b; 95% CI [1429.71618499, 1444.98971204]; votes 6626; rank 69unknownOriginal sourcewinner eligibleThird-party benchmark
#30Gemini 3.5 Flash-LiteGoogle DeepMind1,436.25rating—21,843gemini-3.5-flash-lite; 95% CI [1431.05053629, 1441.44358777]; votes 21843; rank 74unknownOriginal sourcewinner eligibleThird-party benchmark
#31MiniMax-M3MiniMax1,435.20rating—44,544minimax-m3; 95% CI [1430.86827835, 1439.53977389]; votes 44544; rank 77unknownOriginal sourcewinner eligibleThird-party benchmark
#32DeepSeek-V4-FlashDeepSeek1,431.98rating—48,963deepseek-v4-flash; 95% CI [1427.84091263, 1436.10937297]; votes 48963; rank 80unknownOriginal sourcewinner eligibleThird-party benchmark
#33GPT-5.6 LunaOpenAI1,430.14rating—24,433gpt-5.6-luna-xhigh; 95% CI [1425.14784950, 1435.13857729]; votes 24433; rank 82unknownOriginal sourcewinner eligibleThird-party benchmark
#34Mistral Large 3Mistral AI1,427.62rating—65,336mistral-large-3; 95% CI [1424.57601414, 1430.66670291]; votes 65336; rank 85unknownOriginal sourcewinner eligibleThird-party benchmark
#35DeepSeek-V3.2DeepSeek1,424.83rating—46,487deepseek-v3.2; 95% CI [1421.30044809, 1428.36162643]; votes 46487; rank 91unknownOriginal sourcewinner eligibleThird-party benchmark
#36Mistral-Medium-3.5-128BMistral AI1,421.16rating—11,017mistral-medium-3.5; 95% CI [1414.62812191, 1427.69573067]; votes 11017; rank 98unknownOriginal sourcewinner eligibleThird-party benchmark
#37Gemini 2.5 FlashGoogle DeepMind1,417.31rating—122,770gemini-2.5-flash; 95% CI [1414.86923843, 1419.74079237]; votes 122770; rank 111unknownOriginal sourcewinner eligibleThird-party benchmark
#38MiniMax-M2.7MiniMax1,404.90rating—65,411minimax-m2.7; 95% CI [1401.24077059, 1408.56647727]; votes 65411; rank 134unknownOriginal sourcewinner eligibleThird-party benchmark
#39Qwen3.5-35B-A3BQwen1,395.60rating—29,067qwen3.5-35b-a3b; 95% CI [1391.30285568, 1399.90287908]; votes 29067; rank 143unknownOriginal sourcewinner eligibleThird-party benchmark
#40Claude Haiku 4.5Anthropic1,395.25rating—124,979claude-haiku-4-5-20251001; 95% CI [1392.70335168, 1397.78975878]; votes 124979; rank 144unknownOriginal sourcewinner eligibleThird-party benchmark
#41GPT-4.1OpenAI1,382.71rating—49,913gpt-4.1-2025-04-14; 95% CI [1379.00467679, 1386.41941365]; votes 49913; rank 153unknownOriginal sourcewinner eligibleThird-party benchmark
#42DeepSeek-R1DeepSeek1,372.49rating—18,524deepseek-r1; 95% CI [1367.65185984, 1377.33749685]; votes 18524; rank 164unknownOriginal sourcewinner eligibleThird-party benchmark
#43gpt-oss-120bOpenAI1,365.91rating—29,935gpt-oss-120b; 95% CI [1361.54274876, 1370.28718499]; votes 29935; rank 172unknownOriginal sourcewinner eligibleThird-party benchmark
#44Nova 2 LiteAmazon1,362.52rating—12,117nova-2-lite; 95% CI [1356.44886287, 1368.59662424]; votes 12117; rank 178unknownOriginal sourcewinner eligibleThird-party benchmark
#45MiniMax-M2.5MiniMax1,359.16rating—40,895minimax-m2.5; 95% CI [1355.19195179, 1363.12059277]; votes 40895; rank 180unknownOriginal sourcewinner eligibleThird-party benchmark
#46NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA1,348.42rating—15,338nvidia-nemotron-3-nano-30b-a3b-bf16; 95% CI [1342.96803096, 1353.87681829]; votes 15338; rank 191unknownOriginal sourcewinner eligibleThird-party benchmark
#47GPT-4.1 MiniOpenAI1,340.09rating—38,630gpt-4.1-mini-2025-04-14; 95% CI [1335.78819641, 1344.39628941]; votes 38630; rank 195unknownOriginal sourcewinner eligibleThird-party benchmark
#48DeepSeek-V3DeepSeek1,332.28rating—21,770deepseek-v3; 95% CI [1327.55202083, 1337.01331591]; votes 21770; rank 205unknownOriginal sourcewinner eligibleThird-party benchmark
#49Command ACohere1,330.79rating—55,434command-a-03-2025; 95% CI [1327.33626503, 1334.25009199]; votes 55434; rank 208unknownOriginal sourcewinner eligibleThird-party benchmark
#50Olmo-3.1-32B-InstructAi21,311.90rating—11,463olmo-3.1-32b-instruct; 95% CI [1305.73447366, 1318.07100374]; votes 11463; rank 223unknownOriginal sourcewinner eligibleThird-party benchmark
#51Olmo-3-32B-ThinkAi21,299.56rating—5,682olmo-3-32b-think; 95% CI [1291.25214421, 1307.87126458]; votes 5682; rank 230unknownOriginal sourcewinner eligibleThird-party benchmark
#52granite-4.1-8bIBM1,292.07rating—4,027granite-4.1-8b; 95% CI [1281.98658230, 1302.15965212]; votes 4027; rank 237unknownOriginal sourcewinner eligibleThird-party benchmark
#53gpt-oss-20bOpenAI1,287.27rating—10,386gpt-oss-20b; 95% CI [1280.87434855, 1293.66774302]; votes 10386; rank 242unknownOriginal sourcewinner eligibleThird-party benchmark
#54Llama-4-Maverick-17B-128E-InstructMeta1,287.13rating—39,349llama-4-maverick-17b-128e-instruct; 95% CI [1282.87677682, 1291.39199009]; votes 39349; rank 243unknownOriginal sourcewinner eligibleThird-party benchmark
#55GPT-4o MiniOpenAI1,286.22rating—68,698gpt-4o-mini-2024-07-18; 95% CI [1282.70478675, 1289.73790661]; votes 68698; rank 245unknownOriginal sourcewinner eligibleThird-party benchmark
#56GPT-4.1 NanoOpenAI1,284.64rating—6,103gpt-4.1-nano-2025-04-14; 95% CI [1276.99177350, 1292.27849322]; votes 6103; rank 246unknownOriginal sourcewinner eligibleThird-party benchmark
#57GPT-4oOpenAI1,282.33rating—45,499gpt-4o-2024-08-06; 95% CI [1278.14385443, 1286.52509361]; votes 45499; rank 250unknownOriginal sourcewinner eligibleThird-party benchmark
#58Llama-4-Scout-17B-16E-InstructMeta1,279.25rating—29,739llama-4-scout-17b-16e-instruct; 95% CI [1274.53478136, 1283.95578792]; votes 29739; rank 255unknownOriginal sourcewinner eligibleThird-party benchmark
#59Llama-3.3-70B-InstructMeta1,273.91rating—54,411llama-3.3-70b-instruct; 95% CI [1270.42968055, 1277.38585448]; votes 54411; rank 258unknownOriginal sourcewinner eligibleThird-party benchmark
#60Llama-3.1-70B-InstructMeta1,260.70rating—55,240llama-3.1-70b-instruct; 95% CI [1256.98821169, 1264.41220064]; votes 55240; rank 272unknownOriginal sourcewinner eligibleThird-party benchmark
#61phi-4Microsoft1,216.49rating—24,126phi-4; 95% CI [1211.91444642, 1221.07047082]; votes 24126; rank 296unknownOriginal sourcewinner eligibleThird-party benchmark
#62Llama-3.1-8B-InstructMeta1,186.37rating—49,605llama-3.1-8b-instruct; 95% CI [1182.27083365, 1190.47005605]; votes 49605; rank 309unknownOriginal sourcewinner eligibleThird-party benchmark

62 models ranked by the best winner-eligible current-version score when available, otherwise the best publisher-artifact score; highest first. Observed Sep 2, 2026. The model you came from is highlighted.

Methodology and Coverage

Ranking protocolarena rating · rating. Each model appears once; verified winner-eligible runs take precedence, and missing models are not imputed.
Acquisition coverage62 published models from 398 source rows; 333 source identities remain quarantined.

Primary Evidence

Publisher Artifacts

LMArena Vision Arena methodologyLMArena · observed Sep 2, 2026LMArena Text Arena text-2026-09-01-011508720696 score tableLMArena · observed Sep 2, 2026

Questions

LMArena Text Arena FAQs

What does LMArena Text Arena measure?+

Blind pairwise human preferences from LMArena's text arena. Model Markets classifies it as a general language benchmark and preserves the publisher's text-2026-09-01-011508720696 release as a distinct comparison cohort.

How are models ranked on LMArena Text Arena?+

Models are ordered by arena rating in rating, with higher scores ranked first. Each model appears once; a verified winner-eligible current-version result takes precedence over a publisher-artifact-only result, and missing scores are not estimated.

Which model currently leads LMArena Text Arena?+

Claude Opus 5 leads the current verified table with 1,504.96 rating on text-2026-09-01-011508720696. This is a benchmark-specific result, not a universal model-quality claim.

How many models have a published LMArena Text Arena score?+

62 catalog models are published from 398 source rows. 333 source identities remain quarantined rather than guessed.

Can LMArena Text Arena scores be compared with other benchmarks?+

Raw scores should be compared only within the same benchmark version, metric, and protocol. Model Markets normalizes eligible scores only for aggregate rankings, and groups models by an identical benchmark set before ranking them. View aggregate rankings →

Why might a model be missing from LMArena Text Arena?+

A model remains absent when the publisher has no current result, the source model identity is unresolved, the evaluation protocol is incompatible, or the evidence cannot be verified. Model Markets does not substitute a provider claim or infer a score from a related model.

How current is the LMArena Text Arena leaderboard?+

The current Model Markets snapshot was observed Sep 2, 2026 from LMArena artifacts. The exact publisher source and each retained result artifact are linked on this page.

Send Feedback