Unverified50% confidencePredictionExact time
The team is developing Swift 3.8 Flash Next, which has already achieved -30% thinking tokens while maintaining xhigh accuracy
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/14/2026
First Seen
Valid until: 12/13/2026
Sources
Related Entities
Related Claims
UnverifiedQwen3.8 Flash Next is a Flash model using the next-generation Qwen4 architecture, and the complete weights were released on the day of its open-sourcing.66% similarUnverifiedFrom 3.7 Flash to 3.8 Flash, significant improvements have been achieved in three major areas: software engineering capabilities, agentic tasks, and multi-step reasoning.66% similarUnverifiedUnder a dual-Spark TP2 configuration, Qwen 3.8 Flash achieves an inference speed of 55–70 tokens/s and supports 4 fully concurrent agents running simultaneously.63% similarUnverifiedQwen3.8 Flash Next additionally introduces an Ngram module with 51 billion parameters, resembling a massive local phrase memory table.61% similarUnverifiedAlibaba's Tongyi Qianwen officially released the Qwen3.8-Flash-Next (Qianwen 308 Flash NEX) model, with a total parameter count of 181.5B.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/921001API
curl https://kongchang.com/api/v1/knowledge/claims/921001MCP
get_claim(id=921001)