turintech/ kev-4b-artemis:latest

65 Downloads Updated 4 hours ago

Text-only decision-model Kev-4B GGUF files. Default: q4_k_m. TurinTech models, built with Artemis.

ollama run turintech/kev-4b-artemis

curl http://localhost:11434/api/chat \
  -d '{
    "model": "turintech/kev-4b-artemis",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='turintech/kev-4b-artemis',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'turintech/kev-4b-artemis',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated yesterday

yesterday

6b2bd1ea3b1a · 2.5GB

model

archqwen35

·

parameters4.21B

·

quantizationQ4_K_M

2.5GB

license

Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR US

11kB

license

Kev-4B Artemis GGUF Copyright 2026 Turing Intelligence Technology Limited (TurinTech AI) These files

1.0kB

params

{ "num_batch": 4096, "num_ctx": 16384 }

35B

Readme

Kev-4B Artemis

Text-only GGUF files of Kev-4B, a typed decision model. It returns probabilities for allowed answers through /v1/systemone; plain chat is not supported.

TurinTech models, built with Artemis

Ollama

echo "My Stripe integration keeps failing. Please help ASAP." | ollama run turintech/kev-4b-artemis:q4_k_m --ask "Does this message express urgency?"

latest selects q4_k_m. The explicit tags below use the same model bytes as the corresponding Hugging Face files. Existing size tags remain available.

Tag	Hugging Face file	Size
`q4_k_m` (default)	`Kev-4B-Artemis-Q4_K_M.gguf`	2.541 GB
`q3_k_m`	`Kev-4B-Artemis-Q3_K_M.gguf`	2.039 GB
`iq3_m`	`Kev-4B-Artemis-IQ3_M.gguf`	2.165 GB

GB is download size, not runtime memory. Quant labels are size classes for these mixed files; the model card lists tensor types.

For code, send model, state and typed questions to /v1/systemone. Keep questions about one input in the same request; splitting them can change the probabilities.

The package sets num_ctx 16384 and num_batch 4096, the card’s llama-server settings (-c 16384 -ub 4096 -b 4096). The batch size moves the probabilities slightly, so keep it when you compare with the card. Inputs longer than 4,096 tokens are accepted.

Measurements and limitations

The selection panels helped choose the files and do not establish performance on unseen inputs. Same-answer rates measure agreement with the original, including its mistakes. Public files can perform better; see the full comparisons on the card. Measurements used llama-server, not Ollama, and Ollama memory use was not measured.

Full measurements and test conditions.

Licence

Apache-2.0 (licence). See the Hugging Face repository for LICENSE and NOTICE.

These files keep improving.

# Kev-4B Artemis

Text-only GGUF files of Kev-4B, a typed decision model. It returns probabilities for allowed answers through `/v1/systemone`; plain chat is not supported.

TurinTech models, built with Artemis

## Ollama

```bash
echo "My Stripe integration keeps failing. Please help ASAP." | ollama run turintech/kev-4b-artemis:q4_k_m --ask "Does this message express urgency?"
```

`latest` selects `q4_k_m`. The explicit tags below use the same model bytes as the corresponding Hugging Face files. Existing size tags remain available.

| Tag | Hugging Face file | Size |
| --- | --- | --- |
| `q4_k_m` (default) | `Kev-4B-Artemis-Q4_K_M.gguf` | 2.541 GB |
| `q3_k_m` | `Kev-4B-Artemis-Q3_K_M.gguf` | 2.039 GB |
| `iq3_m` | `Kev-4B-Artemis-IQ3_M.gguf` | 2.165 GB |

GB is download size, not runtime memory. Quant labels are size classes for these mixed files; the model card lists tensor types.

For code, send `model`, `state` and typed `questions` to `/v1/systemone`. Keep questions about one input in the same request; splitting them can change the probabilities.

The package sets `num_ctx 16384` and `num_batch 4096`, the card's llama-server settings (`-c 16384 -ub 4096 -b 4096`). The batch size moves the probabilities slightly, so keep it when you compare with the card. Inputs longer than 4,096 tokens are accepted.

## Measurements and limitations

The selection panels helped choose the files and do not establish performance on unseen inputs. Same-answer rates measure agreement with the original, including its mistakes. Public files can perform better; see the full comparisons on the card. Measurements used llama-server, not Ollama, and Ollama memory use was not measured.

[Full measurements and test conditions](https://huggingface.co/turintech/Kev-4B-Artemis-GGUF).

## Licence

Apache-2.0 ([licence](https://huggingface.co/jaredpalmer/kev-4b)). See the Hugging Face repository for `LICENSE` and `NOTICE`.

These files keep improving.

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)