Unverified50% confidenceBenchmarkExact time
In prime-rl, Qwen3 MoE training step speed improved by up to 34%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/29/2026
First Seen
Valid until: 1/27/2027
Sources
Related Entities
Related Claims
UnverifiedUsing QLoRA on A100 and RTX PRO 6000, training speed for Qwen3.5-35B-A3B can be improved by up to 4.1×, and for Qwen3-30B-A3B by up to 3.3×.69% similarUnverifiedThe 3.11x is kernel-level forward acceleration, while 34% is end-to-end training step speed improvement; the two are measured on different bases.67% similarUnverifiedOn Qwen3-8B, after the second round of the loop selected 3 new skills, model performance improved from 64.3% to 66.3%.63% similarUnverifiedOn the NL2RepoBench benchmark, QwenGyre raised the score from 52.5% to 58.5% in 48 steps, and end-to-end training time was shortened by about 44% compared with the Async scheme.62% similarUnverifiedCompared with Qwen3-30B-A3B, Hy-Memory-30B-A3B improves by an average of 11 points, and still scores 56.89 on the EverMemOS framework, which it never saw in training.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1112884API
curl https://kongchang.com/api/v1/knowledge/claims/1112884MCP
get_claim(id=1112884)