arch_center_111/MORPHLING-A4.5-Thinking-Preview

arch_center_111/ MORPHLING-A4.5-Thinking-Preview:latest

22 Downloads Updated 7 months ago

This's the preview for the extended thinking. Soon we will release the final model, which has the Medium and Pro version. Stay tune!

ollama run arch_center_111/MORPHLING-A4.5-Thinking-Preview

curl http://localhost:11434/api/chat \
  -d '{
    "model": "arch_center_111/MORPHLING-A4.5-Thinking-Preview",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='arch_center_111/MORPHLING-A4.5-Thinking-Preview',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'arch_center_111/MORPHLING-A4.5-Thinking-Preview',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 7 months ago

7 months ago

5ed28f7b14b4 · 17GB ·

model

archglm4

parameters32.6B

quantizationQ3_K_L

17GB

template

[gMASK]<sop>{{ if .System }}<|system|> {{ .System }}{{ end }}{{ if .Prompt }}<|user|> {{ .Prompt }}{

142B

params

{ "num_ctx": 30000, "repeat_penalty": 1, "stop": [ "<|system|>", "<|user

158B

Readme

Domain	Benchmark	Metric	MORPHLING-4.5-Thinking-Preview	QwQ-32B-Preview
Reasoning	LiveCodeBench (25.01 – 25.05)	Pass@1 (Thinking)	44.8	43.5
	GPQA Diamond	Accuracy (Thinking)	66.4	65.2%
	AIME 2024	Accuracy (Thinking)	81.9	79.5%
	MATH-500	Accuracy (Thinking)	92.9	90.6%