webAI's TwIL-8B Tops Math and Code Benchmarks at 5 GiB

This story is from 2026-11-05. It is preserved in the archive; the latest stories are on the live feed.

webAI post-trained IBM's Granite 4.2 8B into a reasoning model called TwIL-8B that scores 98.2 on MATH-500 and matches frontier models on code benchmarks while fitting in 5 GiB.

Read the full story at AlphaSignal ↗

Timeline · 1 report

2026-11-05 12:17 · AlphaSignal
webAI's TwIL-8B Beats Frontier Models on Math and Code at 5 GiB

More stories

IBM's Granite-TTM-R3 Runs 18,000 Forecasts per Second Without a GPU — AlphaSignal
MIT study shows AI fixing AI, alongside Mistral open model and Google Genie 4 — Mixture of Experts (IBM)
AI uncovers major web security flaw as new voice model emerges — Mixture of Experts (IBM)
Election backlash against data centers rattles AI stocks and expansion plans — TechCrunch AI
Anthropic strikes $3.4bn deal for gas-fired generators to power AI data centres — Financial Times AI
OpenAI to keep expanding AI data centers despite election results favoring restrictions — DigiTimes Asia
Tencent AI chief warns AI will soon master biology, raising new safety risks — SCMP China Tech
Hydra-PSIE expands 'Coeziune OM–IA' DOI registry to 39 published works — DEV Community — AI

Get the daily brief of stories like this at 6:30 every morning →