Nace.AI Drex v1.5 vs Nace.AI Drex v2.0 (35B-A3B)
Nace.AI Drex v2.0 (35B-A3B) is more accurate on the Jev Decision Index, Nace.AI Drex v2.0 (35B-A3B) is better calibrated, Nace.AI Drex v1.5 is cheaper per 1,000 decisions. Choose on whichever of accuracy, calibration and cost matters most for your decisions.
| Spec | Nace.AI Drex v1.5 | Nace.AI Drex v2.0 (35B-A3B) |
|---|---|---|
| Provider | Nace.AI | Nace.AI |
| Decision Index score | 45.8 | 52.8 |
| Calibration error (ECE) | 0.090 | 0.018 |
| Median latency (p50) | 54 ms | 276 ms |
| Cost per 1,000 decisions | $0.0080 | $0.012 |
| Parameters | 9B | 35B |
| Open weights | Yes | Yes |
| 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 | Decoder LM with a decision readout |
| License | nace-ai-open-rail-m | nace-ai-open-rail-m |
FAQ
Is Nace.AI Drex v1.5 or Nace.AI Drex v2.0 (35B-A3B) more accurate?
Nace.AI Drex v2.0 (35B-A3B) scores higher on the Jev Decision Index (52.8 vs 45.8).
Which is better calibrated, Nace.AI Drex v1.5 or Nace.AI Drex v2.0 (35B-A3B)?
Nace.AI Drex v2.0 (35B-A3B) has the lower expected calibration error (0.018 vs 0.090), so its stated confidence is closer to how often it is right.
Is Nace.AI Drex v1.5 or Nace.AI Drex v2.0 (35B-A3B) cheaper per 1,000 decisions?
Nace.AI Drex v1.5 is cheaper: about $0.0080 vs $0.012 per 1,000 decisions at 200 input tokens each.
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.