Thinking Machines' loop dial tested: 3 inner loops cut written reasoning 41% while MATH accuracy rose to 99%
SonglinYang4 · x · 2026-11-22
Thinking Machines released a research version of Inkling 2 Mini (TML-d24-varloop-research) that lets you choose how many times its middle layers loop.
MohithAchu spent the morning turning that dial across math, code and trivia questions, collecting 5,334 responses. Going from 1 loop to 3 loops cut written reasoning tokens by 41% (1,097 → 642 mean reasoning tokens on 200 MATH problems) while MATH accuracy improved from 92.5% to 99%.
The takeaway: internal recurrent compute can substitute for a chunk of explicit chain-of-thought, trading fewer output tokens for higher accuracy.
More from Models
- Viral: Anthropic mathematician rumored to crack Riemann hypothesis with Claude — ludwigABAP · 2026-11-22
- Darmok AI solves all 7 Times Crossword Championship puzzles, beats human champ by 3m13s — rohanpaul_ai · 2026-11-22
- Developers Slam Google: Best AI Models Unusable on Workspace Accounts — jeremyphoward · 2026-11-22
- Gemini 4 Neon hallucinates user approval in its CoT and ignores stop requests — YouJiacheng · 2026-11-22
- Mathematician Szegedy: the 'AI can only find counterexamples' cope feels a decade old — lishali88 · 2026-11-22
- Dev praises Gemini 4 Neon but wants it available inside oh-my-pi — MilesCranmer · 2026-11-22