Qwen4-Max
AvailableAlibaba's Qwen4 flagship, generally released Nov 16 2026 after the Token Plan-only qwen4-max-preview (Nov 11). Sparse MoE with 2.4T total parameters and 71B active per token (25% fewer active than Qwen3.8-Max at the same scale), with upgraded attention, gated residuals and N-gram embeddings. Accepts text, image and video input and returns text, with a 1M-token context, reasoning on by default, tool calling, structured outputs and web search. Available on QwenCloud API and Token Plan, QwenWork and QwenStudio; $1.50 / $4.50 per 1M input/output tokens ($0.075 cached input). Proprietary; Alibaba says smaller open-weight Qwen4 models will follow. Benchmarks vendor-reported.
Specifications
- License
- Proprietary
- Weights
- Not released
- Architecture
- Mixture-of-Experts
- Parameters
- 2.4T · 71B active
- Context window
- 1M tokens
- Max output
- —
- Knowledge cutoff
- —
- Price (in / out, $/M)
- $1.5 / $4.5
- Modalities
- TextVisionVideoCode
Benchmarks
No benchmark scores recorded yet. Spotted some? Submit a correction.
Vendor-reported figures are claims until independently verified. See methodology.