Qwen3.5 397B A17B
Qwen3.5 397B A17B is Alibaba's flagship model that launched the Qwen3.5 series, unifying separate vision, text, and reasoning model lines into a single model.
Model Overview
Capabilities, design details, and architectural traits
Qwen3.5 397B A17B - The Flagship Model That Launched the Qwen3.5 Series
Qwen3.5 397B A17B is documented as a multimodal foundation model featuring a Hybrid Mixture-of-Experts architecture with early fusion vision-language training. It is the first model Qwen released in the Qwen3.5 series, built to fold capabilities that used to live in separate model lines into one.
| Trait | Detail |
|---|---|
| Defining purpose | A multimodal foundation model with early fusion vision-language training, designed for chat, retrieval-augmented generation, vision-language understanding, video understanding, and agentic workflows. |
| First Qwen3.5 release | Documented as the first model released in the Qwen3.5 series, serving as the flagship entry point for the generation rather than a smaller companion size. |
| Unifies vision and text lines | Built as one native multimodal model, rather than maintaining separate vision focused and text only model lines as in the prior Qwen3 generation. |
| One model, two reasoning modes | Supports both a thinking mode and a non-thinking mode inside a single model, rather than shipping separate instruct and thinking variants as before. |
Built on Gated Delta Networks and Sparse Mixture-of-Experts
Qwen documents the model's efficiency as coming from Gated Delta Networks combined with a sparse Mixture-of-Experts design, aimed at high throughput inference with reduced latency and overhead.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Alibaba | Anthropic | Anthropic |
| Release Date | February 16, 2026 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | - | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 262K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.60 Best Input Pricing | $4 | $2 |
| Output Pricing | $3.60 Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | textimagevideo | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 21.4 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | - | - | - |
| Agentic Index | - | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from Alibaba
Other models by Alibaba
Qwen3.8 Max
Qwen3.8 2.4T A95B
Qwen3.8-Flash-Next
Qwen3.8 27B
Top AI Models
Leading alternatives by intelligence score
