Introducing Qwen1.5
Qwen
Alibaba released Qwen1.5, an open-source language model series available in eight sizes ranging from 0.5B to 110B parameters, integrated with major AI frameworks and platforms. The largest model, Qwen1.5-72B, achieved an 8.61 average score on MT-Bench and outperforms Llama2-70B across all evaluated benchmarks including MMLU, reasoning, and code generation. The release enables developers to access competitive open-source models with 32K token context support through transformers, local inference tools, and API services globally.
Why it matters
GITHUB HUGGING FACE MODELSCOPE DEMO DISCORD Introduction In recent months, our focus has been on developing a “good” model while optimizing the developer experience. As we progress towards Qwen1.5, the next iteration in our Qwen series, this update arrives just before the Chinese New Year. With Qwen1.5, we are open-sourcing base and chat models across six sizes: 0.5B, 1.8B, 4B, 7B, 14B, 32B, 72B, and 110B, and also an MoE model (see blog for more information).