bvassie/qwen3.8:27b-mtp

bvassie/ qwen3.8:27b-mtp

44 Downloads Updated 10 hours ago

Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks.

vision tools thinking 27b

ollama run bvassie/qwen3.8:27b-mtp

curl http://localhost:11434/api/chat \
  -d '{
    "model": "bvassie/qwen3.8:27b-mtp",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='bvassie/qwen3.8:27b-mtp',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'bvassie/qwen3.8:27b-mtp',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated yesterday

yesterday

2a77c382dc8a · 21GB ·

model

archqwen35

parameters26.9B

quantizationQ4_K_M

19GB

projector

archclip

parameters461M

quantizationQ8_0

629MB

draft

GGUF��(��general.architecture��qwen35��general.type��

1.7GB

params

{ "draft_num_predict": 4, "min_p": 0, "presence_penalty": 0, "repeat_penalty": 1,

114B

template

13B

Readme

Created using:

https://huggingface.co/ggml-org/Qwen3.8-27B-GGUF

Modelfile: qwen3.8:27b

FROM ./Qwen3.8-27B-Q4_K_M.gguf
FROM ./mmproj-Qwen3.8-27B-Q8_0.gguf
TEMPLATE {{ .Prompt }}
RENDERER qwen3.8
PARSER qwen3.5
PARAMETER min_p 0
PARAMETER presence_penalty 0
PARAMETER repeat_penalty 1
PARAMETER temperature 1
PARAMETER top_k 20
PARAMETER top_p 0.95

Modelfile: qwen3.8:27b-mtp

FROM ./Qwen3.8-27B-Q4_K_M.gguf
FROM ./mmproj-Qwen3.8-27B-Q8_0.gguf
DRAFT ./mtp-Qwen3.8-27B-Q4_0.gguf
TEMPLATE {{ .Prompt }}
RENDERER qwen3.8
PARSER qwen3.5
PARAMETER min_p 0
PARAMETER presence_penalty 0
PARAMETER repeat_penalty 1
PARAMETER temperature 1
PARAMETER top_k 20
PARAMETER top_p 0.95
PARAMETER draft_num_predict 4