Decider 31B v2 (NVFP4) — VRAM & /v1/systemone setup
Written by Jakub Rusinowski · Last updated
Mapika's 31B decision model on Gemma 4 31B, v2 of the fine-tune, reading text and images: NVFP4 weights and a bf16 vision tower, about 33 GB, for NVIDIA Blackwell GPUs via decider.serve_vllm.
Decider 31B v2 (NVFP4) needs about 21 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
Call Decider 31B v2 (NVFP4)
Served by decider.serve_vllm. This model does not run in Ollama.
- Question types
- choice noul (yes / no) score
- Inputs
- text image
- Serve with
decider.serve_vllm- Endpoint
/v1/systemone
pip install "decider-ai>=1.10.0"Setup for servers other than Ollama →
Build a request for this model Decision models guide
Related guides: Servers other than Ollama · Confidence, thresholds and calibration
Other decision models: Nimble 9B · Tev1 4B · Tev1 0.8B · Winnow 12B · Winnow E4B · Decider 2B · Decider 4B · Decider 35B-A3B (NVFP4) · Decider 12B Vision · JevK5 4B · Intern-Decision 4B · AutoJev 27B · Laya · Clef 27B · Clef-Flash 9B · Rune 26B-A4B v3 · OpenDecider-nano · OpenDecider-small · OpenDecider-small MLX 4-bit · Bosun v3.1 0.6B · Bosun v3.1 1.7B · d1-3B · d1-omni-600M · THX-01 322M · mammoth-1 27B · elephant-1 9B · Strands Decider 2B (hobson-v21) · kev 4B · kev 9B · OpenJev 27B
Hardware fit
Weights plus overhead plus the KV cache at a 8,192-token prompt, on NVIDIA RTX 4090 (24 GB). A publisher build is sized from its file; the other rows are modelled at a standard quant. Decision requests are short, so no long-context figure is shown. Image tokens are not modelled: each image adds to the prompt, so a request with images needs more memory than these rows.
| Quant | Memory | VRAM | Fit |
|---|---|---|---|
| NVFP4 + BF16 vision Publisher build · 32.63 GB file | 35.3 GB | Offload | |
| Q4_K_M 4.83 bpw · modelled quant | 22.4 GB | Tight | |
| Q6_K 6.56 bpw · modelled quant | 29.4 GB | Offload | |
| Q8_0 8.50 bpw · modelled quant | 37.4 GB | Offload |
Published file size: NVFP4 + BF16 vision 32.63 GB. A download size from the model publisher — not a VRAM requirement.
NVFP4 weights with a bf16 vision tower; NVFP4 needs an NVIDIA Blackwell GPU.
Specifications
Verified — Checked against the primary source — the model card or the vendor spec page — and corroborated by a second independent source.
- Parameters
- 32.68B
- Context window
- Not published
- Architecture
- Fine-tune of google/gemma-4-31B-it
- Provider
- Mapika
- Licence
- Apache-2.0
- Availability
- Weights downloadable
- Practical deployment
- Single consumer GPU
- Specified at
- Q4_K_M
- System RAM
- 42 GB
- Record updated
- 2026-10-25
Commercial use permitted. No usage restrictions beyond attribution.
Other Decider sizes
Decider 31B v2 (NVFP4) — frequently asked questions
What is Decider 31B v2 (NVFP4)?
Decider 31B v2 (NVFP4) is a decision model: you send it a state and typed questions (choice, yes/no/unknown, or a score) and it returns one answer per question with a probability for every option. It is not a chat model.
How do I run Decider 31B v2 (NVFP4) locally?
Served by decider.serve_vllm. This model does not run in Ollama. Start it with: `pip install "decider-ai>=1.10.0"`.
How much memory does Decider 31B v2 (NVFP4) need?
About 21 GB for the weights plus overhead at Q4_K_M, before the prompt's KV cache. Decision prompts are short, so the cache stays small.