Unverified50% confidenceFactExact time
Qwen-2引入了GQA、支持128K token的上下文窗口和基于字节级BPE的分词器（词表约15万），参数规模从0.5B到72B
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/7/2026
First Seen
Valid until: 11/5/2026
Sources
阿里通义冲上Text Arena第二名，Qwen模型实力解析
Related Claims
VerifiedThe Qwen series models support context windows ranging from 8K to 128K.71% similarUnverifiedQwen3 在训练时支持长达 128K tokens 的上下文69% similarUnverifiedEarly LLaMA-2 only supported a 4K token context, whereas new-generation models such as DeepSeek-V3 and Qwen2.5 have extended the context to 128K or even longer.68% similarUnverifiedQwen3-27B was officially released with a quantized model in GGUF format, enabling local deployment with a minimum of 8GB of VRAM.67% similarUnverifiedThe Qwen3.6 27B parameter version supports a maximum context window of 128K.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/701171API
curl https://kongchang.com/api/v1/knowledge/claims/701171MCP
get_claim(id=701171)