MODELS
Qwen 2.5
Open-weight general-purpose LLM with strong multilingual and coding chops at low cost.
Specs
- Context window
- 128,000
- Max output
- 8,192
- Modalities
- text
- Tool use
- ✓
- Vision
- —
- Streaming
- ✓
- License
- apache-2.0
- Released
- 2024-09-19
Pricing
- Input / 1M
- $0.27
- Output / 1M
- $0.81
Cost estimate
Qwen 2.5 is Alibaba's open-weight (Apache 2.0) general-purpose LLM, released September 2024 with a 128K context window and tool-calling support. It handles English, Chinese and a wide range of other languages, and the family includes specialized Coder and Math variants. At roughly $0.27/$0.81 per million tokens via API, it targets builders who want a capable, cheap, self-hostable alternative to GPT-4o-class closed models.
Editor's verdict
Pick Qwen 2.5 when you need an open, self-hostable model with solid Chinese and multilingual coverage, or when API cost matters more than absolute reasoning ceiling. It clearly beats Llama 3.1 on Chinese tasks and is competitive on code, but on hard reasoning and agentic workflows it still trails GPT-4o, Claude 3.5 Sonnet and the newer Qwen3 line. Output cap of 8K tokens is also tight for long generations.
Reviews
No reviews yet. Be the first.
Last updated: 2026-04-29