Qwen3.5 397B A17B
by Qwentext+image+video->text10 endpoints
Neura Intelligence Index
66.9/ 100
Rank
#104—
Confidence
mediumBenchmarks
6(15% coverage)
Overview
The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...
Capabilities
Text generation
Image understanding
Video understanding
Tool use / Function calling
Structured output
Extended reasoning
Prompt caching
Long context
Top-P sampling
Stop sequences
Deterministic seed
Log probabilities
Modalities
Input
TextImageVideo
Output
Text
Technical Specifications
Context Window
262.1K tokens
Max Output
235.9K tokens
Tokenizer
Qwen3
Uptime (24h)
9996.9%
Parameters
397000000000B
License
apache_2_0Open Weights
Country
🇨🇳CN
Release Date
February 16, 20267mo
Supported Parameters
frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p
Benchmark Performance
general
| Benchmark | Score | Source |
|---|---|---|
| MMMLU | 88.5 | verified |
coding
| Benchmark | Score | Source |
|---|---|---|
| SWE-bench Verified | 76.4 | verified |
reasoning
| Benchmark | Score | Source |
|---|---|---|
| GPQA Diamond | 88.4 | verified |
| Humanity's Last Exam | 28.7 | verified |
agent
| Benchmark | Score | Source |
|---|---|---|
| BrowseComp | 69 | verified |
| Toolathlon | 38.3 | verified |
Performance Over Time
Pricing
Per 1M tokens
Input$0.55
Output$3.50
Cache read$0.23
Added
February 16, 2026
Last synced 9/18/2026