gutsy-0.8b
Omar (kouhxp)'s CPU decision model: a Qwen3.5-0.8B fine-tune shipped as GGUF for llama.cpp that answers yes/no, choice and score questions with a probability per option plus a 'none of these' reject probability, served by a local Jev-style HTTP runtime.
A full fine-tune of Qwen3.5-0.8B (embeddings frozen) on about 105,000 decision questions, with a cross-entropy plus Brier loss. Ships Q8_0 and Q4_K_M GGUF files with per-type temperature files and the gutsy-inference runtime, which serves /v1/systemone locally, caches the state and shortlists lists of more than 16 options. The maker's own runs on JevBench's public items (a third-party benchmark): 0.732 with Q8_0, 0.697 with Q4_K_M; calibration error is 0.022 on the maker's validation split but 0.107 and 0.137 on JevBench's original and hard tiers. Weak on date and number arithmetic, long multi-step policies and world knowledge. English only.
What it decides
- choice — picks one option from a set
- score — places the input on an ordered scale
- noul — answers a yes/no question with one probability
- classify — assigns a category from a fixed taxonomy
- route — sends the input to one of several destinations
At a glance
| Parameters | 0.8B |
| Base model | Qwen/Qwen3.5-0.8B |
| Maker | Omar (kouhxp) |
| Released | 2026-10-02 |
| License | apache-2.0 |
| Reported accuracy | 73.2% |
Get the weights
pip install systemonemodels
systemone pull kouhxp/gutsy
The files are served from the maker's Hugging Face repository, kouhxp/gutsy, and verified against the checksums recorded here.
Read more
This page was opened by System One for Omar (kouhxp), who can claim the organisation and take it over at any time.