AutoJev 27B — VRAM & /v1/systemone setup
Written by Jakub Rusinowski · Last updated
A 27B multimodal decision model on Qwen3.8-27B, the highest-scoring downloadable model on the Decision Index as of 28 Sep 2026; served by autojev-serve.
AutoJev 27B needs about 18 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 AutoJev 27B
Served by autojev-serve (default port 8000). This model does not run in Ollama.
http://localhost:8000/v1/systemone
Setup for servers other than Ollama →
Build a request for this model Decision models guide
Other decision models: Nimble 9B · Tev1 4B · Tev1 0.8B · Winnow 12B · Winnow E4B · Decider 2B · Decider 4B · Decider 35B-A3B (NVFP4) · JevK5 4B · Intern-Decision 4B · Laya
Affiliate links — we may earn a commission if you sign up, at no extra cost to you.
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.
| Quant | Memory | VRAM | Fit |
|---|---|---|---|
| BF16 (49 GiB) Publisher build · 52.6 GB file | 55.1 GB | Offload | |
| Q4_K_M 4.83 bpw · modelled quant | 19.3 GB | Fits | |
| Q6_K 6.56 bpw · modelled quant | 25.3 GB | Offload | |
| Q8_0 8.50 bpw · modelled quant | 32.1 GB | Offload |
Published file size: BF16 (49 GiB) 52.6 GB. A download size from the model publisher — not a VRAM requirement.
uv run autojev-serve → playground on :8000, POST /v1/systemone with optional base64 images; set AUTOJEV_API_KEY before exposing.
How it was scored
Two different suites on two different scales. Never compare the numbers across the two cards.
Decision Index 0.2.1 (snapshot 2026-09-28)
Decision Index 0.2.1- ECE (lower is better)
- 0.0178
- Median compute
- 101.4 ms on 1x NVIDIA RTX PRO 6000 (96 GB)
Community-maintained; not affiliated with the model authors.
Source ↗Decision models are not ranked on chat, creative or coding scores.
Specifications
Verified — Checked against the primary source — the model card or the vendor spec page — and corroborated by a second independent source. Still unconfirmed: weightsGbByQuant.
- Parameters
- 27.78B
- Context window
- Not published
- Architecture
- Fine-tune of Qwen/Qwen3.8-27B
- Provider
- denis-pplx
- Licence
- Apache-2.0
- Specified at
- Q4_K_M
- System RAM
- 36 GB
- Record updated
- 2026-09-30
Commercial use permitted. No usage restrictions beyond attribution.
AutoJev 27B — frequently asked questions
What is AutoJev 27B?
AutoJev 27B 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 AutoJev 27B locally?
AutoJev 27B does not run in Ollama. Served by autojev-serve (default port 8000). This model does not run in Ollama. See the setup guide for servers other than Ollama.
How much memory does AutoJev 27B need?
About 18 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.
How accurate is AutoJev 27B?
It has two separate published scores on two different suites — the author's own benchmark and the community Decision Index 0.2.1. They are not comparable with each other, and neither is a calibration guarantee.