Skip to content
IFAI

Model rankings

1260 models · 5 boards · 2 sources

WebDev arena · People compare web pages built by two models · Published June 5, 2026

Best web builder
claude-opus-4-7-thinkingAnthropic1,567Score
Best open model
glm-5.1Z.ai1,532Score
Best from China
qwen3.7-max-20260517Alibaba1,537Score
Most voted
claude-haiku-4-5-20251001Anthropic21,389Votes
Top three by Score
1claude-opus-4-7-thinkingAnthropic1,567
2claude-opus-4-8-thinkingAnthropic1,552
3claude-opus-4-6-thinkingAnthropic1,543
Top three open models
1glm-5.1Z.ai · #5 overall1,532
2kimi-k2.6Moonshot AI · #8 overall1,516
3mimo-v2.5-proXiaomi · #16 overall1,466
This week's changes
Fun-Realtime-TTSSpeech · #1New
claude-opus-4-8-thinkingCoding · #2New
minimax-m3Coding · #6New
Krea 2 MediumImage · #7New
RankChangeModelScore95% intervalVotesPrice $/M
1—claude-opus-4-7-thinkingClosedAnthropic1,567±96,234—
2—claude-opus-4-7ClosedAnthropic1,557±95,788—
3Newclaude-opus-4-8-thinkingClosedAnthropic1,552±161,578—
4Newclaude-opus-4-8ClosedAnthropic1,545±142,011—
5▼ 2claude-opus-4-6-thinkingClosedAnthropic1,543±78,881—
6▼ 1claude-opus-4-6ClosedAnthropic1,538±79,812—
7▼ 3qwen3.7-max-20260517ClosedAlibaba1,537±132,466—
8Newminimax-m3ClosedMiniMax1,528±161,680$0.52
9▼ 3claude-sonnet-4-6ClosedAnthropic1,522±712,101—
10▼ 3muse-sparkClosedMeta1,508±161,629—
11▼ 3gemini-3.5-flashClosedGoogle1,506±132,212—
12▼ 3gpt-5.5-xhigh (codex-harness)ClosedOpenAI1,502±95,067—
13▼ 3claude-opus-4-5-20251101-thinking-32kClosedAnthropic1,491±713,064—
14▼ 3qwen3.6-max-previewClosedAlibaba1,484±122,502—
15▼ 3gpt-5.5-high (codex-harness)ClosedOpenAI1,482±95,261—
16▼ 3claude-opus-4-5-20251101ClosedAnthropic1,467±615,302—
17▼ 3qwen3.6-plusClosedAlibaba1,463±87,014—
18▼ 3gpt-5.4-high (codex-harness)ClosedOpenAI1,457±171,482—
19▼ 2gpt-5.5 (codex-harness)ClosedOpenAI1,448±95,125—
20▼ 4gemini-3.1-pro-previewClosedGoogle1,447±611,246—
21▼ 3gemini-3-proClosedGoogle1,439±717,159—
22▼ 3gpt-5.4-medium (codex-harness)ClosedOpenAI1,437±161,449—
23▼ 3gemini-3-flashClosedGoogle1,437±713,276—
24▼ 3mimo-v2-proClosedXiaomi1,433±86,788—
25▼ 2gpt-5.3-codex (codex-harness)ClosedOpenAI1,407±122,963—
26▼ 3gpt-5.2ClosedOpenAI1,405±171,457—
27▼ 3gpt-5.4-mini-highClosedOpenAI1,398±86,400—
28▼ 2gpt-5-mediumClosedOpenAI1,394±133,755—
29▼ 4grok-4.20-beta-0309-reasoningClosedxAI1,393±78,159—
30▼ 3gpt-5.1-mediumClosedOpenAI1,391±96,118—
31▼ 3gpt-5.4ClosedOpenAI1,391±35280—
32▼ 3claude-sonnet-4-5-20250929-thinking-32kClosedAnthropic1,388±715,744—
33▼ 3gemini-3-flash (thinking-minimal)ClosedGoogle1,388±617,358—
34▼ 2claude-opus-4-1-20250805ClosedAnthropic1,386±98,567—
35▼ 4claude-sonnet-4-5-20250929ClosedAnthropic1,386±618,387—
36▼ 2gpt-5.3-codex (codex-harness)ClosedOpenAI1,373±113,549—
37▼ 4grok-4.3ClosedxAI1,371±104,410—
38▼ 3gpt-5.1ClosedOpenAI1,340±712,868—
39▼ 3gpt-5.2-codexClosedOpenAI1,335±87,767—
40▼ 3gpt-5.1-codexClosedOpenAI1,330±106,218—
41▼ 3claude-haiku-4-5-20251001ClosedAnthropic1,324±521,389$2.0
42▼ 3KAT-Coder-Pro-V1Closed1,259±161,881—
43▼ 3gemini-3.1-flash-lite-previewClosedGoogle1,249±710,289$0.56
44▼ 3gpt-5.1-codex-miniClosedOpenAI1,240±171,444—
45▼ 3qwen3.5-flashClosedAlibaba1,238±171,562—
46▼ 3grok-4-1-fast-reasoningClosedxAI1,234±96,912—
47▼ 3grok-4.1-thinkingClosedxAI1,209±201,209—
48▼ 3gemini-2.5-proClosedGoogle1,204±133,297—
49▼ 3mercury-2ClosedInception1,165±23947—
50▼ 3grok-4-fast-reasoningClosedxAI1,150±23934—

Intelligence over time

Each maker's best model at the time on the Artificial Analysis Intelligence Index, by the month it was released.

Gained most · Anthropic +57.2Highest now · Anthropic 64.6Gap between highest and lowest 18.1 ptsTime range Nov 2025 – Nov 2026
64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi

Arena data comes from the leaderboard dataset it publishes on Hugging Face (CC BY 4.0). Only each board's overall leaderboard is used; scores and ranks are unchanged. dataset · CC BY 4.0

Artificial Analysis data comes from its API (artificialanalysis.ai). Scores are unchanged; ranks follow the scores. artificialanalysis.ai

Open or closed follows the license Arena gives; prices and Artificial Analysis intelligence show only where model names match, otherwise “—”.