Unverified50% confidenceBenchmarkExact time
Under the condition of matched per-step generation budgets, FOCUS at the Qwen3.5-9B scale improves by 7.2 percentage points over the strongest baseline (the higher of GRPO/VinePPO) on the BrowseComp-Plus benchmark, reaching 55.4 versus 48.2.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/27/2026
First Seen
Valid until: 1/25/2027
Sources
Related Entities
Related Claims
UnverifiedVASPO improves average accuracy by 2.59 percentage points over PPO on Qwen3-4B and by 1.81 percentage points on OLMo3-7B.73% similarUnverifiedOn the Qwen3-VL series of models (from 2B to 30B-A3B) and on InternVL3.5-8B, SPLIT-RL improves average accuracy by 1.4 to 6.1 percentage points compared to standard GRPO.70% similarUnverifiedWhen run on Qwen3.6-35B-A3B with a 50% expert pruning ratio, SlimWise increases decoding throughput by up to 1.81x with minimal accuracy loss.68% similarUnverifiedCompared to the previous generation Qwen3.6-27B, Qwen3.8-27B shows significant improvements on multiple benchmarks, with Deep SWE rising from 13.3 to 42.2 (nearly tripling) and OS World rising from 63.9 to 84.3.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1102724API
curl https://kongchang.com/api/v1/knowledge/claims/1102724MCP
get_claim(id=1102724)