1.7B Model Outperforms Qwen3-8B in Strict Formal Logic Reasoning
Creative-Fig522 · reddit · 2026-08-22
A Reddit user highlighted TwIL-LM2, a 1.7B parameter model that achieved surprising results in strict formal logic reasoning, beating larger competitors like Qwen3-8B and Gemma-4-26B.
- Specs: TwIL-LM2 is a PEFT LoRA adapter based on SmolLM2-1.7B, specialized purely for formal logic translation.
- Benchmarks: On the strict-7 scoring (no partial credit, exact format required), it scored 0.2386, ahead of Qwen3-8B (0.2093) and Gemma-4-26B (0.2050). However, larger models still win on the loose-match average.
- Implication: This challenges the narrative that reasoning strictly requires massive scale. If high efficiency can be achieved via narrow specialization on small models (1-3B), it suggests a future pipeline of specialists rather than routing everything to a monolithic 70B model.
- License: Non-commercial license.
Related event: Tiny 1.7B TwIL-LM2 Beats Larger Models at Formal Logic Reasoning(2 posts)→
More from Models
- Muse Spark Open runs locally on a Mac M3 Ultra via llama.cpp: frontier model on desktop — pcuenq · 2026-10-16
- OpenAI ships GPT-Live-1.1: new voices in six languages, fewer confirmations — OpenAI · 2026-10-16
- Muse Spark model appears on LMArena, testable in Battle and Agent modes — arena · 2026-10-16
- Running a multi-model UI design arena: Paseo orchestrates four design drafts per model — vista8 · 2026-10-16
- Ilya Sutskever questions Meta license clause taxing MaaS providers above $20M revenue — ilyasu · 2026-10-16
- Muse-Spark open-weights drop: 816B-A42B MoE model gets 359GB MLX 3-bit build — ns123abc · 2026-10-16