Unverified50% confidenceBenchmarkExact time
By fine-tuning Qwen3.5-4B on approximately 3,000 browser operation trajectories, a developer improved accuracy on the BrowseWebApp Bench benchmark from 22% to 63%.
1
Sources
50%
Confidence
Long-term
Relevance
9/7/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedQwen3 27B achieved a 100% pass rate in the 256K long-context memory retrieval test (needle in a haystack), with all 15 runs being accurate.70% similarUnverifiedQwen3.8 27B achieved an overall score of 52 on the Artificial Analysis evaluation platform.68% similarUnverifiedThe fine-tuned Qwen3.5-4B model improved accuracy on BU Bench V1 from 15% to 53%.67% similarUnverified基准测试显示tiktoken比HuggingFace的tokenizers库快3-6倍67% similarUnverifiedQwen3 27B在Kanban看板应用测试中消耗近87%上下文，一次生成出无bug的成果，支持拖拽、排序、筛选、归档等功能66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/869642API
curl https://kongchang.com/api/v1/knowledge/claims/869642MCP
get_claim(id=869642)