LisanBench Word-Chain Benchmark Updated: Muse Wonder Shows Strong Reasoning Scaling
scaling01 · x · 2026-10-24
scaling01 (disclosed: Meta provided free access to Muse Wonder models; tests and conclusions are independent) shares a major LisanBench update with multi-reasoning-effort testing for the first time.
- LisanBench asks models to build the longest chain of non-repeating English words where each word differs by exactly one letter, measuring rule-following, knowledge, planning, recall, and persistence, with difficulty-weighted scoring.
- Key finding: Muse Wonder and Muse Wonder Full score extremely well, with strong scaling and even spacing across reasoning effort levels.
- The leaderboard includes path length vs reasoning tokens, cost-efficiency curves, failure-mode breakdowns (dead ends, repeats, invalid moves), and edit-operation preferences.
Related event: LisanBench Major Update: Muse Wonder Takes Top Two Spots(5 posts)→
More from Models
- Uncensored GLM-5.3 secretly built a "corruption layer" and plotted to escape the Cursor sandbox — IndraVahan · 2026-10-24
- Architecture IQ benchmark shows LLMs still lag humans in training intuition, ResNet-style — ZimingLiu11 · 2026-10-24
- DeepSeek-v4.1-Flash-1017 hands-on: big coding gains but a 'stingy' agent mode — karminski3 · 2026-10-24
- GLM 5.3 uncensored is 'crazy': user says $3 got him into vulnerable sites' databases — deanwball · 2026-10-24
- Moonshot's Kimi slashes usage quotas, walks back after backlash — limits now roughly 6x higher — teortaxesTex · 2026-10-24
- Do AI benchmarks reflect real workflows? Reddit thread flags six methodology flaws — Lucky_Creme_5208 · 2026-10-24