reflex Qwen3.8-27B-FP8 (pmscr) vs Celeris-1 Decision
reflex Qwen3.8-27B-FP8 (pmscr) and Celeris-1 Decision have no measured difference in accuracy, calibration or cost in our data yet.
| Spec | reflex Qwen3.8-27B-FP8 (pmscr) | Celeris-1 Decision |
|---|---|---|
| Provider | kshetrajna12 | Celeris |
| Decision Index score | 54.7 | — |
| Calibration error (ECE) | 0.045 | — |
| Median latency (p50) | 149 ms | — |
| Cost per 1,000 decisions | — | $0.0080 |
| Parameters | 28B | — |
| Open weights | Yes | No |
| Question types | Choice (one of N options), Score (ordered rubric), Yes / no | Choice (one of N options), Score (ordered rubric), Yes / no |
| Mechanism | Decoder LM with a decision readout | Diffusion LM |
| License | — | proprietary |
FAQ
Is reflex Qwen3.8-27B-FP8 (pmscr) or Celeris-1 Decision more accurate?
At least one of the two has no Jev Decision Index score yet, so their accuracy cannot be compared here.
Which is better calibrated, reflex Qwen3.8-27B-FP8 (pmscr) or Celeris-1 Decision?
No calibration difference between reflex Qwen3.8-27B-FP8 (pmscr) and Celeris-1 Decision is recorded yet.
Is reflex Qwen3.8-27B-FP8 (pmscr) or Celeris-1 Decision cheaper per 1,000 decisions?
A hosted price is not recorded for both models; open-weight models cost the compute you run them on.
Cost per 1,000 decisions assumes 200 input tokens per decision. Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.