Unverified50% confidenceFactExact time
The Qwen3 series training data reportedly exceeds 36 trillion tokens, covering code, mathematics, science, and multilingual text.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedQwen3覆盖从0.6B到235B的多种参数规模，支持思考模式（thinking mode）混合推理能力，覆盖119种语言的多语言预训练数据78% similarUnverifiedThe reviewer believes that Qwen3.8 27B is stronger than the previous generation Qwen3.7 in almost all aspects of coding capability.68% similarUnverifiedThe Qwen series comes in multiple parameter sizes ranging from 1.5 billion to 72 billion.65% similarUnverifiedCorpus 3.6 27B是开源社区基于阿里千问3.6（Qwen-3.6）进行后训练得到的优化版本，继承了千问3.6的256K长上下文能力65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/915553API
curl https://kongchang.com/api/v1/knowledge/claims/915553MCP
get_claim(id=915553)