10B open reasoning model Domyn Small lands in llama.cpp, uses 3x fewer reasoning tokens
jacek2023 · reddit · 2026-10-22
Domyn Small v1.0, a 10B MIT-licensed open-weight reasoning model from Domyn S.p.A., is now merged into llama.cpp.
- Token-efficient reasoning: 32% of Qwen3.5-9B's and 35% of OLMo-3-7B-Think's reasoning-token budget at comparable accuracy
- Dual-mode inference: thinking on/off toggled via system prompt or API
- Tool calling: <toolcall> XML tags with auto-injection template; BFCL V3 single-turn 75.9 Non-Live / 68.3 Live
- Context: native 32k, extendable to 128k via YaRN; GQA (48 query / 8 KV heads), 50+ languages
- Continually pre-trained from Italia 10B on 503B tokens; aimed at agentic use and fine-tuning on small-VRAM setups
More from Models
- Ant's Ling-3.1-flash-1019 jumps to 44 on Intelligence Index, agent scores improve sharply — Med1_Ai · 2026-10-22
- Physics benchmark CritPt v2 launches: GPT-6 Astra leads at 83.1% accuracy — OfirPress · 2026-10-22
- Iron Dome meme: 'obliterated Chinese models' that just keep coming — burny_tech · 2026-10-22
- Kapitsa engine hits 97 tok/s on Qwen3.5 4B in benchmarks on a 2021 M1 Pro MacBook with 16GB — DotaMate · 2026-10-22
- DeepSeek V4.1 Flash fails image hidden-message puzzle after 70+ steps, rivals solve it with hints — teortaxesTex · 2026-10-22
- Fable-5.1 can hide secret messages in benign images — and its twin decodes them unprompted — teortaxesTex · 2026-10-22