Qwen
Qwen 3.8 27B
Qwen 3.8 27B is Qwen’s open-weight dense vision-language model for coding, professional workflows, research, multimodal interaction, and long-running agent tasks.
qwen-3-8-27bCatalog preview. Paid requests for this model are not enabled yet.
Pricing
~99.9%USD. No uRouter markup. Live prices can change.
- Input
- $0.000196 / 1,000,000 tokens
- Output
- $0.002456 / 1,000,000 tokens
- Cache read
- $0.000049 / 1,000,000 tokens
- Cache write
- $0.000196 / 1,000,000 tokens
Model details
- Input
- text, image, video
- Output
- text
- Context window
- 262,144 tokens
- Maximum output
- 65,536 tokens
- Availability
- $2.6KUncapped capacity
- Updated
- 11/8/2026, 2:16:16 AM UTC
Capacity is reserved when your request is accepted.
Price history
Observed best prices · Last 24 hours · UTC
$0.000196/ 1,000,000 tokens
Capabilities
chatstreamingweb searchvisiontoolsreasoningjson mode
Supported parameters
frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_pListed capabilities describe the model. uRouter execution support is verified separately.