The best AI models right now

Who's #1 for everything, from millions of blind head-to-head votes on Arena (updated Oct 9, 2026). Bars show how far each model trails the leader.

Best AI model

👑 Gemini 4 Argon

Google DeepMind · 1534 Elo

1. 2Claude Opus 5.5−19
2. 3Claude Fable 5.1−24
3. 4Claude Opus 5−27
4. 5Claude Opus 4.6−31

Overall text Arena · Arena, Oct 8

Best for coding

👑 Claude Opus 5.5

Anthropic · 1813 Elo

1. 2GPT-6 Astra−27
2. 3Claude Sonnet 5.5−39
3. 4GPT-6.1 Sol−58
4. 5Claude Fable 5.1−69

WebDev Arena · Arena, Oct 8

Best agent

👑 Claude Opus 5.5

Anthropic · 14.3 score

1. 2GPT-6 Astra−1.2
2. 3Claude Fable 5.1−1.6
3. 4Claude Sonnet 5.5−2.3
4. 5GPT-6.1 Sol−2.6

Agent Arena · Arena, Oct 8

Best open model

X

👑 MiMo V2.6 Pro

xiaomi · 1492 Elo

1. 2MKimi K3−16
2. 3ZGLM 5.3−21
3. 4ZGLM 5.3 Flash−21
4. 5ZGLM 5.2−22

Overall text, open weights only · Arena, Oct 8

Polymarket odds

Who will be #1 next?

All odds →

Which company has best AI model end of 2026?

Google43%

Anthropic42%

OpenAI10%

Meta3%

$2.4M vol · ends Dec 31Polymarket ↗

Which company has the best AI model end of October?

Google▼569%

Anthropic▲531%

OpenAI<1%

Meta<1%

$2.2M vol · ends Oct 31Polymarket ↗

Which company has the best AI model end of November?

Google▼949%

Anthropic▲1040%

OpenAI3%

Meta3%

$267.2K vol · ends Dec 1Polymarket ↗

Specialist leaderboards

Best at everything else

Best for hard problems

👑 Gemini 4 Argon

Google DeepMind · 1548 Elo

1. 2Claude Opus 5.5−13
2. 3Claude Opus 4.6−21
3. 4Claude Opus 5−22

Hard prompts · Arena, Oct 8

Best at math

👑 Gemini 4 Argon

Google DeepMind · 1536 Elo

1. 2Claude Opus 5−5
2. 3Gemini 3.8 Flash−8
3. 4Claude Fable 5.1−11

Math prompts · Arena, Oct 8

Best writer

👑 Claude Opus 5.5

Anthropic · 1533 Elo

1. 2Gemini 4 Argon−2
2. 3Claude Fable 5.1−26
3. 4Claude Opus 4.6−28

Creative writing · Arena, Oct 8

Best AI search

👑 GPT-5.6 Sol

OpenAI · 1257 Elo

1. 2Claude Opus 4.6 Search−4
2. 3GPT-5.5 Search−15
3. 4Claude Opus 4.7−24

Search Arena · Arena, Aug 24

Best vision

👑 Claude Fable 5

Anthropic · 1324 Elo

1. 2Claude Opus 5.5−3
2. 3Claude Opus 5−5
3. 4Claude Fable 5.1−6

Vision Arena · Arena, Oct 9

Best image model

👑 GPT Image 2.5 Sunburst

OpenAI · 1425 Elo

1. 2GPT Image 2.5 Flare−27
2. 3GPT Image 2−42
3. 4Grok Imagine Image 2.0 (canvas)−91

Text-to-image Arena · Arena, Oct 7

Best video model

👑 Gemini Omni 1.1 Flash

Google DeepMind · 1516 Elo

1. 2Gemini Omni Flash−3
2. 3BFlux 3 Video−23
3. 4Grok Imagine Video 1.5 Agent−24

Text-to-video Arena · Arena, Sep 22

Source: Arena (arena.ai) public leaderboards. Variants of one model (high, max, thinking) are shown once at their best score. Agent board scores are win rates, not Elo.