reflex Qwen3.8-27B-FP8 (pmscr)
reflex Qwen3.8-27B-FP8 (pmscr) by kshetrajna12 is a decision model: it answers typed questions with a probability for each allowed answer instead of generating text. It scores 54.7 on the Jev Decision Index (rank 35).
Pricing
No hosted price is listed. The weights are open: the cost is your own compute.
Specifications
- Provider
- kshetrajna12
- Parameters
- 28B
- Released
- Sep 2026
- Open weights
- Yes
- Question types
- Choice (one of N options), Score (ordered rubric), Yes / no
- Mechanism
- Decoder LM with a decision readout
- Backbone
- Qwen/Qwen3.8-27B-FP8
- Jev-compatible request
- No
- Runs on CPU
- No
- License
- —
Benchmarks
- Jev Decision IndexECE 0.045
- 54.7
- Jev Decision Indexedition 0.2.1 (earlier edition)
- 52.2
- DecisionBenchECE 0.070
- 0.578
Jev Decision Index: public 57.3, ECE 0.045, Brier 0.374. Lower ECE means its probabilities can be taken at face value.
Self-reported by the publisher
The model's own claims, often on its own sets or a self-run copy of a public suite. Not independently verified and never ranked alongside measured results.
- Jev Decision Index public suite (self-run)Self-reportedDecision Index public balanced skill, edition 0.3 kit (release-v3), self-run
- 57.3
Speed & latency
Compare reflex Qwen3.8-27B-FP8 (pmscr) with…
FAQ
How much does reflex Qwen3.8-27B-FP8 (pmscr) cost?
reflex Qwen3.8-27B-FP8 (pmscr) has no hosted price listed; its weights are open, so the cost is the compute you run it on.
What questions can reflex Qwen3.8-27B-FP8 (pmscr) answer?
reflex Qwen3.8-27B-FP8 (pmscr) answers choice (one of n options), score (ordered rubric), yes / no questions, returning a probability for each allowed answer.
Is reflex Qwen3.8-27B-FP8 (pmscr) well calibrated?
On the Jev Decision Index its expected calibration error (ECE) is 0.045; below about 0.05, a stated 80% confidence is right roughly 80% of the time.
Is reflex Qwen3.8-27B-FP8 (pmscr) open weight, and can it run on a CPU?
Yes, reflex Qwen3.8-27B-FP8 (pmscr) has open weights. It needs a GPU to run at useful speed.
Data as of 2026-10-14. Source: Metatext decision-model catalogue, 2026-10-14 (Jev Decision Index, DecisionBench). Pricing is per 1M tokens (USD); confirm with the provider before production use.