Compare Start here

Alibaba QwenLanguage model

Qwen3.8-Max

Alibaba Qwen language model.

Compare head-to-head Race it in 3D

Overall Intelligence

#14

of 61
21.7 pts behind #1

2

races led

18

weeks at #1

69.4

score

Released: Aug 3, 2026; 3 mo ago
Context: 1M; 131K max output
Price: $2 / $6; per 1M in/out · $3.00 blended · OpenRouter
Speed: 38 tok/s; 1.95s to first token (AA)
Modalities: text · image · video; outputs text
Weights: Closed; API and apps

The 30-second version

Alibaba Qwen language model.

Key takeaways

1Alibaba Qwen language model.

Where it ranks

Latest real results · click any race to watch it

Qwen3.8-MaxBest on each axis

Overall

14Overall+21.7 pts

Coding

17Coding+40.1 pts 10Frontend & UI+133 Elo 18Coding Q&A+45 Elo 18Frontier SWE+47.7 pts 8React Web Apps+163 Elo 24Agentic Coding+18.8 pts 14Terminal & DevOps+24.7 pts

Hard Reasoning

28Frontier Questions+18.3 pts

Math

9Math+14.6 pts 10Competition Math±0.0 pts 17Research Math+18.9 pts 12Math Chat+35 Elo

Science

18Science+28.9 pts 19Graduate Science QA+3.7 pts 8Medicine & Health+53 Elo 35Scientific Coding+13.7 pts 7Science Q&A+36 Elo

Writing & Language

11Writing & Language+18.2 pts 13Creative Writing+55 Elo 12Professional Writing+56 Elo 17Multilingual+50 Elo 15Instruction Following+56 Elo 12Conversation+69 Elo

Multimodal

9Multimodal+14.9 pts 9Image Understanding+12 Elo

Tool Use & Agents

1Tool Use & Agents99.1 1Tool Use51.3 2Web Search±0 Elo

Long Context

32Long Context+44.2 pts 32Long-Document Reasoning+8.4 pts

Speed & Cost

26Speed & Cost+61.3 pts 30Output Speed+352 tok/s 16Time to First Token+1.46s 27Intelligence per Dollar+603.1 pts

Trust & Factuality

35Trust & Factuality+63.8 pts 31Factuality+28.7 pts

Overall Intelligence over time

#1 frontierQwen3.8-MaxGPT-6 AstraOpus 4.7

Update history

The distilled version first; open any release for the deeper breakdown.

The Qwen Max family

Qwen3April 2025 Qwen3-MaxSeptember 2025 Qwen3.6-MaxApril 2026 Qwen3.8-MaxAugust 2026 3.8-Max PrimeSeptember 2026