Decider 35B-A3B
Decider 35B-A3B by Mapika is a decision model: it answers typed questions with a probability for each allowed answer instead of generating text. It scores 48.5 on the Jev Decision Index (rank 49).
Pricing
No hosted price is listed. The weights are open: the cost is your own compute.
Specifications
- Provider
- Mapika
- Parameters
- 36B
- 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/Qwen3.5-35B-A3B-Base
- Jev-compatible request
- Yes
- Runs on CPU
- No
- License
- apache-2.0
- Hugging Face
- Mapika/decider-35b-a3b
Benchmarks
- Jev Decision IndexECE 0.023
- 48.5
- DecisionBenchECE 0.044
- 0.561
Jev Decision Index: public 47.7, ECE 0.023, Brier 0.388. Lower ECE means its probabilities can be taken at face value.
Speed & latency
Compare Decider 35B-A3B with…
FAQ
How much does Decider 35B-A3B cost?
Decider 35B-A3B has no hosted price listed; its weights are open, so the cost is the compute you run it on.
What questions can Decider 35B-A3B answer?
Decider 35B-A3B answers choice (one of n options), score (ordered rubric), yes / no questions, returning a probability for each allowed answer.
Is Decider 35B-A3B well calibrated?
On the Jev Decision Index its expected calibration error (ECE) is 0.023; below about 0.05, a stated 80% confidence is right roughly 80% of the time.
Is Decider 35B-A3B open weight, and can it run on a CPU?
Yes, Decider 35B-A3B 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, Enterprise Decision Arena, Benchmark Heaven JevBench). Pricing is per 1M tokens (USD); confirm with the provider before production use.