qwentext → text
Qwen: Qwen3 235B A22B Instruct 2507
Canonical ID: qwen/qwen3-235b-a22b-2507
Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass. It is optimized for general-purpose text generation, including instruction following,...
Artificial Analysis Benchmarks
VerifiedNo standardized Artificial Analysis benchmark scores recorded for this variant yet.
Token Pricing & Economics
Prompt Input (1M Tokens)
Input$0.14
Completion Output (1M Tokens)
Output$0.80
Prompt KV-Cache Read (1M Tokens)
Cached$0.05
Context Window Capacity
Context256k tokens
Supported Parameters & Features
frequency_penaltylogit_biaslogprobsmax_tokensmin_ppresence_penaltyrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p