Unverified50% confidenceBenchmarkExact time
Osprey在Llama-3.3-70B-Instruct上平均接受长度提升21.2%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
UnverifiedOsprey在Qwen3-8B上平均接受长度提升16.1%74% similarUnverifiedDGM在SWE-bench软件工程基准测试中自动将基线性能提升了一倍以上,实现了30个百分点的绝对提升62% similarUnverified在完全相同的模型下,仅靠Harness策略优化,ARC-AGI-3得分从13.3%提升到38.3%,同时输出Token节省约六倍57% similarUnverifiedDOS 3 Note采用新的强化学习方案Tempo,在ARC-AGI 3基准上得分超过30%57% similarUnverified相比上一代Qwen3.6-27B,Qwen3.8-27B在多项测试上大幅提升,其中Deep SWE从13.3提升到42.2(接近三倍),OS World从63.9提升到84.357% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/893136API
curl https://kongchang.com/api/v1/knowledge/claims/893136MCP
get_claim(id=893136)