Tess-4.5-27B released: open fine-tune beats its Qwen base using 30% fewer reasoning tokens
udmrzn · x · 2026-10-20
Developer migtissera released Tess-4.5-27B, an open fine-tune of Qwen3.8-27B.
- Surpasses its base model on engineering evaluations while using 30% fewer reasoning tokens
- Fine-tuned on real engineering sessions generated with Claude Fable 5, Fable 5.1, and Opus 5.5
- Reasoning traces were generated with a constitution to preserve intelligence density at reduced token counts
Weights, GGUF quants, and a blog post are all live on Hugging Face.
Related event: Tess-4.5-27B Open-Sourced, Beats Qwen Base With Fewer Inference Tokens(2 posts)→
More from Models
- Mistral's Voxtral 3 speech models land #3 on STT leaderboard at $13/1M chars — ArtificialAnlys · 2026-10-21
- Perplexity ships Decider V1.2, tops Vals AI's new ClaimBench decision-model benchmark — bo_wangbo · 2026-10-21
- Omnii-1 genome model debuts: first AI-designed phages ever given to a human patient — JosephJacks_ · 2026-10-21
- Nous Portal Pricing Revealed: Unused Credits Roll Over, Up to $100 Cap on $200 Tier — alexcovo_eth · 2026-10-21
- 1.3B open-source doc parser OXR tops three benchmarks, parsing 100 pages in 101 seconds — s_mohinii · 2026-10-21
- How LLMs store facts: three learning phases, attention circuits, and hallucinations — scott_linderman · 2026-10-21