← All models

Qwen 3.5 397B

Alibaba

flagship

Capabilities

VisionToolsCodeExtended ThinkingStructured OutputMultilingual

Strengths

  • Latest Qwen flagship — natively multimodal (text+image+video)
  • 262K context
  • MoE (17B active) — 60% cheaper to run than predecessor
  • LiveCodeBench v6 83.6%, SWE-bench 76.4%
  • 201 languages supported
  • Apache 2.0

Weaknesses

  • Large model
  • Self-hosting impractical for most

Pricing

Input / 1M tokens

$0.385

Output / 1M tokens

$2.4499999999999997

Context window

262k

Transparency

Open weights

10.0 / 10

Open training data

3.0 / 10

Open methodology

7.0 / 10

Licence openness

10.0 / 10

Provider disclosure

2.0 / 10

FMTI company score

300.0 / 10

Composite:6.4 / 10

Apache 2.0, open weights. Alibaba FMTI 30.

Sustainability

Inference energy

5.0 / 10

Training footprint

N/A

Provider infrastructure

N/A

Composite:5.0 / 10

MoE efficient. Depends on hosting.

Task Fitness

question answering
79%
code
85%
mathematics
84%
business communication
64%
analysis
82%
extraction
78%
long-form generation
75%
research
57%
reasoning
82%
summarisation
78%
translation
84%
conversation
75%