webAI Releases TwIL-8B, an 8.8B Reasoning Model Scoring 98.2 on MATH-500
📝 内容摘要
webAI released TwIL-8B, an 8.8B reasoning model post-trained from IBM Granite 4.2 8B. It posts 98.2 on MATH-500 and 81.6
webAI released TwIL-8B, an 8.8B reasoning model post-trained from IBM Granite 4.2 8B. It posts 98.2 on MATH-500 and 81.6 on LiveCodeBench, but trails similar-size models on knowledge benchmarks.
The post webAI Releases TwIL-8B, an 8.8B Reasoning Model Scoring 98.2 on MATH-500 appeared first on MarkTechPost.
打开微信「扫一扫」，分享给好友或朋友圈
关闭