TwIL 8B
An 8.8B reasoning model for math, code and instruction following, post-trained from ibm-granite/granite-4.2-8b with full-parameter supervised fine-tuning and GRPO/DAPO reinforcement learning. It ships in two variants that share one architecture and tokenizer: balanced (the default) and verbose. See Variants. Against open models of similar size and up to 3.5x larger, TwIL-8B posts the highest MATH-500 (98.2) and LiveCodeBench (81.6 †) scores in the comparison and leads Qwen3.5-9B on IFBench (+9.9 †). It is level with Qwen3.5-9B on contest math while using fewer tokens per problem.
TwIL 8B launched on 14 October 2026. It ranked #48 on Daily Launch on 12 November 2026 with a Launch Score of 31, from trending score 225 on Hugging Face. It moved down 2 places from the day before.
What is TwIL 8B?
TwIL 8B is a model by webAI-Official on Hugging Face. It is released under the webai-non-commercial-license-ver.-1.0 license.
How to get TwIL 8B
Get it from its model card on Hugging Face.
Launch timeline
Launch Score, last 7 days
#83 of 228 in the roundup for the week of 9-15 November 2026.