The honest view from inside technical leadership
LLM Leaderboard · General chat (human-judged)

Text Arena leaderboard

Blind head-to-head votes on everyday prompts, from writing and analysis to Q&A, turned into an Elo-style rating.

← All models

Rankings

Snapshot 22 November 2026 · Scores copied from each benchmark's own leaderboard · How the CTO Index works

Score and price

#ModelProviderText ArenaPrice $/M
1Gemini 4 CarbonGoogle1527$2 / $10
2Gemini 4 ArgonGoogle1520$2 / $10
3Gemini 4 NeonGoogle1513$1.50 / $7.50
4Muse WonderMeta1511$4 / $20
5Claude Opus 5.5Anthropic1506$4 / $20
6Claude Fable 5Anthropic1504$10 / $50
7Claude Fable 5.2Anthropic1503$8 / $40
8Claude Fable 5.1Anthropic1503$10 / $50
9Muse Spark 1.3Meta1498$1.25 / $4.25
10Gemini 3.8 FlashGoogle1497$0.75 / $3.75
11Gemini 4 HydrogenGoogle1490$1.50 / $7.50
12Claude Opus 5Anthropic1490$5 / $25
13Gemini 3.7 FlashGoogle1490$0.75 / $3.75
14Kimi K3Moonshot AI1488$3 / $15
15GPT-5.6 SolOpenAI1485$4 / $20
16Qwen3.8-MaxAlibaba1484$2 / $6
17GPT-6.1 SolOpenAI1483$2 / $10
18Kimi K3.1 ProMoonshot AI1482$2 / $10
19GLM-5.3Z.ai1482$1.40 / $4.40
20Claude Opus 4.8Anthropic1482$5 / $25
21MiMo-V2.6-ProXiaomi1481$0.43 / $0.87
22GPT-6 AstraOpenAI1475$10 / $50
23DeepSeek V4.1 FlashDeepSeek1475$0.15 / $0.60
24GLM-5.3-FlashZ.ai1475$0.02 / $0.07
25GPT-6.1 TerraOpenAI1474$1 / $5
26Claude Sonnet 5.5Anthropic1473$2 / $10
27GPT-6.2 AstraOpenAI1468$8 / $40
28GPT-6.2 SolOpenAI1465$2 / $10
29Claude Sonnet 5Anthropic1461$2 / $10
30GPT-6 SolOpenAI1458$2 / $10
31Grok 4.6xAI1457$1.60 / $4.80
32Step 5 PreviewStepFun1449$1 / $2.70
33MiMo-V2.6-FlashXiaomi1448$0.14 / $0.28
34Grok 4.7xAI1443$2 / $6
35GPT-6 LunaOpenAI1443$0.10 / $0.50
36Qwen3.8-27BAlibaba1435$0.40 / $3
37Beam 1.0Reflection AI1434$0.30 / $1
38Mistral Large 4Mistral1423$0.70 / $2.10
39Claude Haiku 5.5Anthropic1421$0.10 / $0.50
40Inkling 2 SmallThinking Machines1412$0.30 / $1

Other benchmarks

Frequently asked questions

What does Text Arena measure?

Blind head-to-head votes on everyday prompts, from writing and analysis to Q&A, turned into an Elo-style rating.

Which model has the highest Text Arena score?

Gemini 4 Carbon from Google leads Text Arena at 1527, ahead of Gemini 4 Argon (1520).

Why does Text Arena matter to engineering leaders?

A good proxy for how the model feels to the rest of the company: drafting documents, summarising, answering questions.

What are the limitations of Text Arena?

Style and formatting sway votes. A 10-point Elo gap is roughly a 51/49 preference split, so the top cluster is effectively tied.