Saif 0.8B Lightning decision model
Saif 0.8B Lightning returns typed probabilities from open noncommercial weights. Read its current public-suite scores and reported timing setup.
By GyrTechnology · v0.3 weight update: 7 Nov 2026
A 0.8B decision model fine-tuned from Qwen3.5-0.8B. It reads a shared state and isolated question segments in one forward pass, returning probabilities for choice, yes/no and score questions. The release includes serving code and weights without a text-generation head. Its non-commercial licence limits how it can be used.
Quick answers
What is Saif 0.8B Lightning by GyrTechnology?
Saif 0.8B Lightning: A 0.8B decision model fine-tuned from Qwen3.5-0.8B.
Who makes Saif 0.8B Lightning?
Saif 0.8B Lightning: GyrTechnology
Are Saif 0.8B Lightning's weights open?
Saif 0.8B Lightning: Public decoder and decision-row weights, Python runtime and Decision Index engine Licence: CC BY-NC 4.0; non-commercial
Facts
- Size
- 0.8B; about 1.5 GB in bf16
- Base
- Qwen/Qwen3.5-0.8B
- Licence
- CC BY-NC 4.0; non-commercial
- Inputs
- Text and JSON state; choice, noul and score questions
- Access
- Public decoder and decision-row weights, Python runtime and Decision Index engine
Claims and sources
- GyrTechnology reports 47.75 chance-corrected skill and 60.73 raw on the public Decision Index 0.3 suite, with all 140,178 requests answered. Median/mean/p80 latency is 18/54/30 ms on one RTX 5070 Ti, one request at a time. These are maker results; the earlier v0.1 scored 42.79. vendor reportedsource
Links
Source: Current card and weights → Source: Release files →
Other models by GyrTechnology
Related projects
No related projects are listed in the registry.