Qwen3-235B-A22B is a 235B parameter mixture-of-experts (MoE) model developed by Qwen, activating 22B parameters per forward pass. It supports seamless switching between a "thinking" mode for complex reasoning, math, and...
Tested on 10 benchmarks · BenchGecko score 55.2. Top scores: Chatbot Arena Elo — Overall (1375.2%), Lech Mazur Writing (83.0%), MATH level 5 (68.9%).
Qwen3 235B A22B Thinking 2507 scores 55.3 (100% as good) at $0.23/1M input · 49% cheaper
Multi-language code editing from Aider. Tests editing ability across Python, JavaScript, TypeScript, Java, C++, Go, Rust, and more.
Unusual and adversarial machine learning challenges. Tests robustness of reasoning about edge cases in ML systems.
Deceptively simple questions that humans find easy but AI models often get wrong. Tests common sense and reasoning gaps.
Competition-level math from AMC, AIME, and olympiad problems. Level 5 is the hardest tier, requiring creative problem-solving.
- Typetext
- Context131K tokens (~66 books)
- ReleasedApr 2025
- LicenseOpen Source
- StatusActive
- Cost / Message~$0.003
Frequently Asked Questions
Key facts · as of 2026-10-05
- Qwen3 235B A22B by Alibaba Qwen. BenchGecko score 55.2, rank 115 of 312 scored models (normalized average of public benchmark scores).
- List price $0.46 input · $1.82 output per 1M tokens (as of 2026-10-05).
- Sold by 1 provider (as of 2026-10-05): Alibaba $0.46 in / $1.82 out.
How to cite · data as of 2026-10-05
Qwen3 235B A22B · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/qwen3-235b-a22b
Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP