Unverified50% confidenceReleaseExact time
OpenBMB发布了面向LLM预训练的开放英文网页语料库UltraFineWeb，总量超过1万亿token、约11.4亿篇文档，全部来自Common Crawl并以Apache 2.0许可证开放
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/24/2026
First Seen
Valid until: 11/22/2026
Sources
Related Entities
Related Claims
VerifiedOpen Code是一个拥有超过十万GitHub星标的开源AI编码代理项目67% similarUnverifiedOpenCode 通过 models.dev 目录可调用来自 Anthropic、Google、OpenAI、智谱（GLM）、DeepSeek 等数十家提供商的模型65% similarUnverifiedOpenCode社区拥有44000+社区Skill64% similarUnverifiedOpenAI的微调服务需要企业上传JSONL格式的训练样本64% similarUnverified完整的OpenUSD库依赖Boost、TBB、OpenSubdiv等数十个第三方库63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/794100API
curl https://kongchang.com/api/v1/knowledge/claims/794100MCP
get_claim(id=794100)