# Kev model sizes and sources

> Kev is Jared Palmer's family of open, Jev-compatible models, from 0.8B to 27B parameters. Kev 1.0 pins checkpoints and serving code.

Updated 2026-11-04 (Australia/Adelaide)

By Jared Palmer · Released: 2026-10-01 · OPEN

[Kev](<https://decisionmodels.ai/models/kev>) is Jared Palmer's family of open, Jev-compatible decision models, from 0.8B to 27B parameters. A pointer head on a Qwen base scores each option, and the API matches TypeSafe's, so their SDK works against a local [Kev](<https://decisionmodels.ai/models/kev>) server. [Kev](<https://decisionmodels.ai/models/kev>) 1.0, released 1 Oct 2026, pins the family's checkpoints, model cards, evaluation suites and serving code; it introduces no newly trained weights.

## Facts

- Sizes: 0.8B, 4B, 9B, 27B
- Licence: Apache-2.0
- Base: Qwen3.5 (27B: Qwen3.8)
- Release: [Kev](<https://decisionmodels.ai/models/kev>) 1.0, 1 Oct 2026
- Stars: 9,758 on GitHub (GitHub API, 2 Nov 2026)
- Price: $0.042 per million input tokens and $0 per million output tokens for jaredpalmer/kev-4b on OpenRouter (VENDOR-REPORTED, 2 Nov 2026) · [Compare prices](<https://decisionmodels.ai/pricing>)
- Validated context: VENDOR-REPORTED: 65,536 tokens for Kev-27B; 8,192 for 0.8B, 4B and 9B, measured on CUAD contracts

## Claims and sources

- **VENDOR-REPORTED**: [Kev](<https://decisionmodels.ai/models/kev>) 1.0's breadth-v1 held-out-dataset test scored a chance-corrected index of 23.3 for 0.8B, 38.0 for 4B, 41.0 for 9B and 52.3 for 27B, against [Jev](<https://decisionmodels.ai/models/jev>)'s 54.0, in the maker's fp32 evaluation at shipped temperatures. No [Kev](<https://decisionmodels.ai/models/kev>) trained on these datasets. [source](<https://github.com/jaredpalmer/kev/releases/tag/kev-1.0>)
- **VENDOR-REPORTED**: Kev-27B's 65,536-token validated context is the longest tested state whose CUAD accuracy stays within three percentage points of its 8k-token accuracy at the 95% lower bound. The smaller models already miss that tolerance at 16k. [source](<https://github.com/jaredpalmer/kev/releases/tag/kev-1.0>)

## Links

- [Source: GitHub](<https://github.com/jaredpalmer/kev>)
- [Source: Kev 1.0 release](<https://github.com/jaredpalmer/kev/releases/tag/kev-1.0>)
- [Source: Hugging Face](<https://huggingface.co/jaredpalmer>)
- [Source: Kev 4B on OpenRouter](<https://openrouter.ai/jaredpalmer/kev-4b>)
- [Source: GitHub API: stars](<https://api.github.com/repos/jaredpalmer/kev>)

## Related projects

### Jev and Kev compared on the same questions

[Jev and Kev compared on the same questions](<https://decisionmodels.ai/projects#project-jev-vs-kev>)

[Source: opper-ai/jev-vs-kev](<https://github.com/opper-ai/jev-vs-kev>)

Benchmark code and a 362-question test set that sends each item to [Jev](<https://decisionmodels.ai/models/jev>) and to [Kev](<https://decisionmodels.ai/models/kev>) 4B, and reports accuracy, calibration, latency and cost.

- Creator: Opper

- Stars: 1

- Licence: Apache-2.0

- Model: [Kev](<https://decisionmodels.ai/models/kev>), [Jev](<https://decisionmodels.ai/models/jev>)

- Tags: benchmark, evaluation

### LinkedIn feed filter

[LinkedIn feed filter](<https://decisionmodels.ai/projects#project-linkedin-focus-filter>)

[Source: GuitarWag/linkedin-focus-filter](<https://github.com/GuitarWag/linkedin-focus-filter>)

A Chrome extension that hides LinkedIn posts that don't match your stated interests, using [Kev](<https://decisionmodels.ai/models/kev>) 0.6B running in the browser on WebGPU. Post text never leaves the browser.

- Creator: GuitarWag

- Stars: 1

- Licence: MIT

- Model: [Kev](<https://decisionmodels.ai/models/kev>)

- Tags: browser, extension