Press Space to continue
Smaller models like TwIL-LM3-Pro (3.66B parameters) are now
Press Space to continue
smaller models are getting insanely smart. this new 3.66B model comes close to the 120B gpt-oss on math word problems: in webAI’s GSM8K evaluation: > TwIL-LM3-Pro: 94.3% > gpt-oss-120b: 97.7% > x33 FEWER PARAMS it’s called TwIL-LM3-Pro, built by WebAI, for checking deductions and critiquing proofs. you don’t need a massive model for every little problem.
open on x · 2 viewsViews and clicks count once per visitor per day. Clicks are opens of this post on X.