Alibaba LLMs
Chinese cloud provider behind the Qwen open-weight model family.
The Qwen series (Qwen2.5, Qwen3, Qwen3-Coder) is one of the most-downloaded open-weight families. Models range from 0.5B to 235B active parameters and are widely used for fine-tuning and self-hosting.
Founded: 1999 · HQ: Hangzhou, China · Docs: www.alibabacloud.com ↗
All Alibaba models
| Model | Family | Context | Input $/M | Output $/M | Released | Status |
|---|---|---|---|---|---|---|
| Qwen3-Coder-480B | Qwen3 | 1.0M | $2.00 | $6.00 | 2025-07-22 | active |
| Qwen3-235B | Qwen3 | 128K | $0.50 | $2.00 | 2025-04-29 | active |
Comparisons with other providers
The most-searched comparisons involving Alibaba models:
- Qwen3-Coder-480B vs GPT-5 Nano
- Qwen3-Coder-480B vs GPT-4o Mini
- Qwen3-235B vs GPT-5 Nano
- Qwen3-235B vs GPT-4o Mini
Working with the Alibaba API
Documentation lives at https://www.alibabacloud.com/help/model-studio. Before paying for any call, count input tokens with the appropriate counter:
- OpenAI Token Counter — for GPT-* and o-series models
- Claude Token Counter — for Anthropic Claude models
- Gemini Token Counter — for Google Gemini models