Alibaba releases Qwen3.8-Flash, previewing Qwen4 architecture
Qwen3.8-Flash is a 125B-parameter multimodal MoE with 262K native context (extensible to 1M via YaRN), priced at $0.16/1M input and $0.47/1M output tokens on QwenCloud. It's an early preview of the Qwen4 architecture, now open-weight.
How this story unfolded
3 days · 2 reports · 5 community posts · from Aug 25
Qwen by email
Get an email when Qwen has news
No news that day, no email.
More stories today
- SentinelOne CEO discusses earnings and AI's cybersecurity impact
- Andrew Ng: Biggest AI opportunities aren't where you think
- OpenAI co-founder warns of closing window to secure internet
- Yann LeCun: Provably safe AI is impossible
- Gemini 3.8 Flash preview reportedly in use at Google