Currently listed through these providers:
Model details
Qwen Max
Qwen Max is built on the Qwen2.5 architecture as a large-scale Mixture of Experts model, designed to deliver the strongest inference performance within the Qwen family. The model is engineered specifically for complex multi-step tasks, making it especially capable when handling layered reasoning or intricate workflows. It also incorporates specialized upgrades in agent programming and tool invocation compared to its preview version, positioning it as a step forward in practical autonomous task execution.
The model was pretrained on over 20 trillion tokens and further refined through Supervised Fine-Tuning and Reinforcement Learning from Human Feedback methodologies. This post-training pipeline helps shape its instruction-following behavior and overall stability. With its agent-focused enhancements and multi-step task strength, Qwen Max fits well into use cases requiring reliable tool use, structured planning, and sustained reasoning across extended contexts.
Quick Info
Powered by- Provider
- Alibaba
- Model key
- qwen-max
- Release date
- Apr 3, 2024
- Last updated
- Jan 25, 2025
- Knowledge cutoff
- 2024-04
- Input modalities
- Output modalities
- Capabilities
Cost
- Input token cost
- $1.60
- Output token cost
- $6.40
Limits
- Output tokens
- 8,192 tokens
- Context window
- 32,768 tokens
Transparent token rates
Compare Qwen Max pricing
Rates are shown per one million tokens. Combined means one million input plus one million output tokens.
Latest news about Qwen Max
No articles yet. Fetch the latest news to show it here.