qwentext + image → text
Qwen: Qwen3 VL 235B A22B Instruct
Canonical ID: qwen/qwen3-vl-235b-a22b-instruct
Qwen3-VL-235B-A22B Instruct is an open-weight multimodal model that unifies strong text generation with visual understanding across images and video. The Instruct model targets general vision-language use (VQA, document parsing, chart/table...
Artificial Analysis Benchmarks
VerifiedNo standardized Artificial Analysis benchmark scores recorded for this variant yet.
Token Pricing & Economics
Prompt Input (1M Tokens)
Input$0.21
Completion Output (1M Tokens)
Output$1.90
Prompt KV-Cache Read (1M Tokens)
Cached$0.10
Context Window Capacity
Context256k tokens
Supported Parameters & Features
frequency_penaltylogit_biaslogprobsmax_tokensmin_ppresence_penaltyrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p