Qwen: Qwen3 235B A22B Thinking 2507

openrouter.ai

Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...

Description

Specs

Type	Model
Section	Models & platforms / text->text
Pricing	paid (от $0.23/mo)
Platform	Self-hosted
Systems	api, python, self-hosted
Hosting	cloud
Site language	en
Vendor	Qwen
Launched	2025-07-25

Can do

Platforms

api python self-hosted

Found in sources

Similar in «Models & platforms»

Cost-efficient open-source MoE model rivaling GPT-4o in reasoning and math tasks

Open reasoning and coding models on par with the top closed ones.

European open and commercial models, Le Chat and Codestral.

Research faster with Sider AI—summarize web pages, take notes, generate reports, and chat with leading AI models. Powered by ChatGPT, Gemini, Claude, Grok, Deepseek, etc. Try the best AI sidebar now!

The hub of models and datasets: millions of weights, Spaces, Transformers.

Передовая мультимодальная AI-платформа от ByteDance с современными языковыми, визуальными и речевыми моделями, а также интегрированными возможностями рассуждения и поиска.