Decision models SOTA: Nace.AI Drex v2.0 (35B-A3B)

CategoryDecision models· Decision models

Also in Decision models:VisionC

Decision models · Oct 14, 2026

Who is SOTA in Decision models?

Nace.AI Drex v2.0 (35B-A3B)

Typed answers with probabilities. The crown follows Decision Index Full and JevBench, both independent at weight 1. Vendor-reported numbers are display-only at weight 0. As of Oct 14, 2026, Nace.AI Drex v2.0 (35B-A3B) is state of the art in Decision models (since Oct 14, 2026; sources Decision Index ECE; methodology v1.0).

Overall Open weights Calibration Latency Price Hosted ≤ 32B Best per $

Lowest ECE among models in the top quartile of accuracy. Brier breaks ties.

Standings · calibration

The board

RSS History →

#	Model	State	Confidence	Why
1	NANace.AI Drex v2.0 (35B-A3B) Nace.AI	SOTA · Decision models	medium	Nace.AI Drex v2.0 (35B-A3B) leads Decision models calibration since 2026-10-14, with the lowest ECE in the top quartile of accuracy.
2	UNKev 27B (Kev 1.0, Qwen3.8-27B) Unknown	Near-SOTA · Decision models	medium	Ranked #2 in Decision models calibration. Kev 27B (Kev 1.0, Qwen3.8-27B) is inside the top quartile of accuracy.
3	TOtorchcast-decision-27b (Torchcast AI, Qwen3.8-27B fine-tune) Torchcast AI	Near-SOTA · Decision models	medium	Ranked #3 in Decision models calibration. torchcast-decision-27b (Torchcast AI, Qwen3.8-27B fine-tune) is inside the top quartile of…
4	OMOneJev1.5-27B OmniJev	#4 in Decision models	medium	Ranked #4 in Decision models calibration. OneJev1.5-27B is inside the top quartile of accuracy.
5	UNDecider 31B (NVFP4) Unknown	#5 in Decision models	medium	Ranked #5 in Decision models calibration. Decider 31B (NVFP4) is inside the top quartile of accuracy.
6	OMOneJev1.5-9B OmniJev	#6 in Decision models	medium	Ranked #6 in Decision models calibration. OneJev1.5-9B is inside the top quartile of accuracy.
7	STSieve 31B Sthānika AI	#7 in Decision models	medium	Ranked #7 in Decision models calibration. Sieve 31B is inside the top quartile of accuracy.
8	FAFastino GLiDE no-thinking (28B) Fastino	#8 in Decision models	medium	Ranked #8 in Decision models calibration. Fastino GLiDE no-thinking (28B) is inside the top quartile of accuracy.
9	PIBlink v0.3 26B-A4B (NVFP4) Pixilab.ai	#9 in Decision models	medium	Ranked #9 in Decision models calibration. Blink v0.3 26B-A4B (NVFP4) is inside the top quartile of accuracy.
10	VLvLLM-SR d3 (27B) vLLM Semantic Router	#10 in Decision models	medium	Ranked #10 in Decision models calibration. vLLM-SR d3 (27B) is inside the top quartile of accuracy.
11	ORWald-26B-A4B ORG2 AI	#11 in Decision models	medium	Ranked #11 in Decision models calibration. Wald-26B-A4B is inside the top quartile of accuracy.
12	UNreflex Qwen3.8-27B-FP8 (pmscr) Unknown	#12 in Decision models	medium	Ranked #12 in Decision models calibration. reflex Qwen3.8-27B-FP8 (pmscr) is inside the top quartile of accuracy.
13	FIDarwin-27B-ZTC FINAL_Bench	#13 in Decision models	medium	Ranked #13 in Decision models calibration. Darwin-27B-ZTC is inside the top quartile of accuracy.
14	UNdecider chat on Gemma-4-31B-it (Mapika, frozen base, inference technique) Unknown	#14 in Decision models	medium	Ranked #14 in Decision models calibration. decider chat on Gemma-4-31B-it (Mapika, frozen base, inference technique) is inside the top…
15	UNAJev lora1 (Gemma 4 26B-A4B) Unknown	#15 in Decision models	medium	Ranked #15 in Decision models calibration. AJev lora1 (Gemma 4 26B-A4B) is inside the top quartile of accuracy.

Embed

Badges for the leaders

Nace.AI Drex v2.0 (35B-A3B)

[![SOTA · Decision models · Oct 2026](https://isitsota.com/badge/drex-v2-0/decision.svg?slice=calibration)](https://isitsota.com/m/drex-v2-0)

Calibration

Reliability · Nace.AI Drex v2.0 (35B-A3B)

Confidence on the horizontal axis, accuracy on the vertical axis. Larger marks carry more weight.

Display only

Self-reported claims

No vendor-reported row is in this snapshot. Vendor-reported numbers are display-only at weight 0.

Best page and recent changes

Best AI model for decision models

What is a decision model

Evidence recipe

What can move this crown

Decision Index Full — primary, weight 1
JevBench — primary, weight 1
Decision Index latency — display, weight 0 (display only, cannot crown)
JevBench price — display, weight 0 (display only, cannot crown)
Vendor decision claims — display, weight 0 (display only, cannot crown)

Who is SOTA in Decision models?

As of Oct 14, 2026, Nace.AI Drex v2.0 (35B-A3B) is state of the art in Decision models (since Oct 14, 2026; sources Decision Index ECE; methodology v1.0).

Since when is the Decision models crown current?

The current crown in Decision models has held since Oct 14, 2026.

What is a decision model?

A decision model takes state, as text or as images, plus a typed question. It returns a typed answer with probabilities rather than free text. Output types: noul (yes/no/unknown), choice, score, multi-label.

How is calibration ranked?

The calibration slice ranks the lowest ECE among models in the top quartile of accuracy, so a constant probability cannot take the slice. Brier breaks ECE ties.

Do vendor-reported decision numbers count?

Vendor-reported numbers are display-only at weight 0.