Text Arena: LLM Rankings & AI Model Benchmarks

Lab Rank		Model Score		Rank Spread
1	Google Gemini 4 Argon (High) · Proprietary	1532±9	1	11
2	Anthropic Claude Opus 5.5 (High) · Proprietary	1516±7	2	25
3	Meta Muse Wonder (Max) · Proprietary	1502±9	6	214
4	Xiaomi MiMo-V2.6-Pro · MIT	1491±9	11	524
5	Alibaba Qwen3.8-Max · Proprietary	1483±5	15	1028
6	Moonshot Kimi K3 (Max) · Kimi K3 license	1476±5	24	1534
7	Z.ai GLM-5.3 (Max) · MIT	1474±5	27	1637
8	OpenAI GPT-5.5 (High) · Proprietary	1471±4	30	2237
9	Baidu ERNIE 5.1 · Proprietary	1468±5	34	2440
10	DeepSeek DeepSeek-V4.1-Flash-20260910 (Max) · MIT	1464±6	38	2548
11	SpaceXAI Grok 4.20 Beta (Reasoning, 0309) · Proprietary	1451±4	51	4269
12	Amazon amazon-nova-experimental-chat (26-02-10) · Proprietary	1448±10	53	3989
13	Bytedance Dola Seed 2.0 Pro · Proprietary	1448±3	54	4474
14	Nvidia NVIDIA-Nemotron-3-Ultra-550B-A55B (NVFP4) · OpenMDW-1.1	1445±7	63	4490
15	Thinky Inkling · Apache 2.0	1442±4	72	5390
16	StepFun Step 5 Preview · Proprietary	1440±7	74	5098
17	Tencent Hy3 · Apache 2.0	1438±6	80	55100
18	MiniMax MiniMax-M3 · MiniMax Community License	1433±4	91	70107
19	Mistral Mistral Large 4 · Proprietary	1431±7	93	70119
20	Meituan LongCat-Flash-Chat Exp (2602) · Proprietary	1427±4	102	87124
21	Reflection Beam 1.0 501B A23B · Apache 2.0	1425±9	104	77138
22	Upstage Solar Pro 4 · Proprietary	1388±5	170	158177
23	IBM Granite 4.2 30B · Apache 2.0	1361±10	197	177210
24	Ant Group Ling-flash-2.0 · MIT	1361±7	198	181209
25	INTELLECT-3 · MIT	1356±8	202	188213
26	Inception AI Mercury 2 · Proprietary	1354±10	203	187216
27	Cohere Command A (03-2025) · CC-BY-NC-4.0	1331±3	226	216235
28	Ai2 Olmo 3.1 32B Instruct · Apache 2.0	1311±6	243	232251
29	Microsoft Phi-4 · MIT	1217±5	315	306317

Lab Rank

Model Score

Rank Spread

Google

Gemini 4 Argon (High) · Proprietary

1532±9

Anthropic

Claude Opus 5.5 (High) · Proprietary

1516±7

Meta

Muse Wonder (Max) · Proprietary

1502±9

214

Xiaomi

MiMo-V2.6-Pro · MIT

1491±9

524

Alibaba

Qwen3.8-Max · Proprietary

1483±5

1028

Moonshot

Kimi K3 (Max) · Kimi K3 license

1476±5

1534

Z.ai

GLM-5.3 (Max) · MIT

1474±5

1637

OpenAI

GPT-5.5 (High) · Proprietary

1471±4

2237

Baidu

ERNIE 5.1 · Proprietary

1468±5

2440

DeepSeek

DeepSeek-V4.1-Flash-20260910 (Max) · MIT

1464±6

2548

SpaceXAI

Grok 4.20 Beta (Reasoning, 0309) · Proprietary

1451±4

4269

Amazon

amazon-nova-experimental-chat (26-02-10) · Proprietary

1448±10

3989

Bytedance

Dola Seed 2.0 Pro · Proprietary

1448±3

4474

Nvidia

NVIDIA-Nemotron-3-Ultra-550B-A55B (NVFP4) · OpenMDW-1.1

1445±7

4490

Thinky

Inkling · Apache 2.0

1442±4

5390

StepFun

Step 5 Preview · Proprietary

1440±7

5098

Tencent

Hy3 · Apache 2.0

1438±6

55100

MiniMax

MiniMax-M3 · MiniMax Community License

1433±4

70107

Mistral

Mistral Large 4 · Proprietary

1431±7

70119

Meituan

LongCat-Flash-Chat Exp (2602) · Proprietary

1427±4

102

87124

Reflection

Beam 1.0 501B A23B · Apache 2.0

1425±9

104

77138

Upstage

Solar Pro 4 · Proprietary

1388±5

170

158177

IBM

Granite 4.2 30B · Apache 2.0

1361±10

197

177210

Ant Group

Ling-flash-2.0 · MIT

1361±7

198

181209

INTELLECT-3 · MIT

1356±8

202

188213

Inception AI

Mercury 2 · Proprietary

1354±10

203

187216

Cohere

Command A (03-2025) · CC-BY-NC-4.0

1331±3

226

216235

Ai2

Olmo 3.1 32B Instruct · Apache 2.0

1311±6

243

232251

Microsoft

Phi-4 · MIT

1217±5

315

306317

Text Arena🏆Overall

View overall rankings across various AI models in text-to-text tasks across math, coding, creative writing, and other open-ended domains.

Remove Style Control Leaderboard Plots

Text Arena🏆Overall

View overall rankings across various AI models in text-to-text tasks across math, coding, creative writing, and other open-ended domains.

Remove Style Control Leaderboard Plots

Fraction of Model A Wins in A vs. B Battles (Ties Excluded)

Average Win Rate Against Each Opponent (Ties Excluded)

Battle Count for Each Pair of Models (Ties Excluded)

Confidence Intervals on Model Strength (via Bootstrapping)