Qwen3.6 35B A3B
Alibaba's open-weight coding model with 35B total and 3B active parameters via mixture-of-experts. Delivers near-frontier coding performance at budget pricing with Apache 2.0 license.
Specifications
| Lab | Qwen |
|---|---|
| Context window | 262.1K tokens |
| Input price | $0.15/1M |
| Output price | $1.00/1M |
| Release | Apr 2026 |
Strengths
- State-of-the-art open-weight coding performance
- Exceptional cost efficiency ($0.15/M input)
- Strong agentic coding with tool use
- Apache 2.0 license for commercial use
- Runs on consumer hardware (3B active params)
Weaknesses
- Below frontier on general reasoning
- Limited outside coding and technical domains
- MoE requires more RAM than active params suggest
Best for
- Agentic coding workflows
- Cost-sensitive coding deployments
- Self-hosted code assistants
- Open-source AI development
In Depth: Qwen3.6 35B A3B
Summary
Qwen3.6 35B A3B is an AI model from Qwen.
Released in Apr 2026. It currently appears in the Overall category on LMRank. It takes text input and produces text output, with a context window of 262.1K tokens. Input pricing is $0.15 per 1M tokens and output is $1.00 per 1M tokens on OpenRouter.