3B TwIL Model Outperforms 120B Open Source Model on Formal Reasoning
Socially-great8275 · reddit · 2026-09-02
webAI released TwIL, a family of formal logic models (1.7B and 3B). The 3B TwIL-LM3 beats the 120B gpt-oss-120b on 4 out of 5 formal reasoning benchmarks while being 40x smaller and runnable on an iPhone. It uses WiSE-FT weight interpolation to mitigate catastrophic forgetting during fine-tuning. The models are released under a non-commercial license.
More from Models
- GPT 5.6 Sol animation claims spark technical debate — chongdashu · 2026-09-02
- Uncensored MiniMax-H3-Turbo LoRA Trends on Hugging Face — Pepe104 · 2026-09-02
- Speculation: Anthropic's Fable Is Opus Looping Twice, Hence the 2x Price — cgarciae88 · 2026-09-02
- Gemini 3.8 Flash spotted rolling out to Pro/Ultra accounts — Able-Line2683 · 2026-09-02
- Fable 5.1 recreates Mario World 1-1; critics call for a level-design benchmark — max_paperclips · 2026-09-02
- Model Scaling Trend: 2T Parameters Becoming New Norm as KV Cache Shrinks 10x YoY — zephyr_z9 · 2026-09-02