Qwen 3 Coder released by Alibaba is making all the news in the AI industry, only leaving 13 days for Kimi K2 by MoonShot to enjoy their spotlight. The coding model is a mixture of expert LLM that is specialized on trillions of tokens that are mostly coding data. It employs pre-training and post-training techniques that seem to show that smaller models can perform as well as SOTA models from OpenAI, Gemini, and Anthropic. The 480B with 35B active parameter model is now available for us to use. Exciting times for Qwen3 Coder and LLM industry!
#ai #qwen #alibaba #llm
