通义听悟
Great for meeting-heavy teams, students, journalists, and researchers turning speech into knowledge; not for formal archival use without human proofreading. Verification level not recorded ·
qwen/qwen3-235b-a22b-thinking-2507
Rank #248
qwen
qwen/qwen3-235b-a22b-thinking-2507
Usage data date 2026-09-12 · updated daily
Usage and specs
Daily rank#248
Daily tokens1.5B
Daily requests265.8K
Daily change
−30.5%
Released
What is qwen/qwen3-235b-a22b-thinking-2507
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...
qwen/qwen3-235b-a22b-thinking-2507 capabilities
Input: Text Output: Text
qwen/qwen3-235b-a22b-thinking-2507 API pricing
Input · per 1M tokens$0.23
Output · per 1M tokens$2.30
USD list-price snapshot, not independently verified; confirm current pricing and terms on the provider site before purchase.
Tools built on this model
通义听悟
Great for meeting-heavy teams, students, journalists, and researchers turning speech into knowledge; not for formal archival use without human proofreading. Verification level not recorded ·