Instructions to use KoarAI/LFM2.5-350M-Thinking-0003 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use KoarAI/LFM2.5-350M-Thinking-0003 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="KoarAI/LFM2.5-350M-Thinking-0003")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("KoarAI/LFM2.5-350M-Thinking-0003")
model = AutoModelForCausalLM.from_pretrained("KoarAI/LFM2.5-350M-Thinking-0003", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use KoarAI/LFM2.5-350M-Thinking-0003 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "KoarAI/LFM2.5-350M-Thinking-0003"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KoarAI/LFM2.5-350M-Thinking-0003",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/KoarAI/LFM2.5-350M-Thinking-0003

SGLang

How to use KoarAI/LFM2.5-350M-Thinking-0003 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "KoarAI/LFM2.5-350M-Thinking-0003" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KoarAI/LFM2.5-350M-Thinking-0003",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "KoarAI/LFM2.5-350M-Thinking-0003" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KoarAI/LFM2.5-350M-Thinking-0003",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use KoarAI/LFM2.5-350M-Thinking-0003 with Docker Model Runner:
```
docker model run hf.co/KoarAI/LFM2.5-350M-Thinking-0003
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

🐨 KoarAI / LFM2.5-350M-Thinking-0003

📌 Release Note: Model Revision `0003`

Model Revision: 0003
Key Architecture & Dataset Improvements:

Anti-Overfitting Training Policy: Trained with a calibrated 2.3 epochs limit with a cosine learning rate scheduler (lr=2.5e-5), preventing catastrophic forgetting and repetition loops while preserving pristine generation fluency.

Multi-Teacher Distillation Mix (2,831 hand-crafted samples):

r0b0tlab/qwen3.8-max-distillation-50k (quality_score >= 6): Deep mathematical, algorithmic, and code reasoning distilled directly from Qwen 3.8 Max (700 samples).

Vikhrmodels/GrandMaster-PRO-MAX: High-quality natural Russian conversational mastery and instruction following (800 samples).

iamtarun/python_code_instructions_18k_alpaca: Clean Python algorithms, web development, and code syntax generation (600 samples).

open-thoughts/OpenThoughts-114k: Step-by-step verification and multi-hop Chain-of-Thought traces encapsulated in <think> blocks (431 samples).

TIGER-Lab/MMLU-Pro: Complex STEM, business, logic, and law benchmark multiple-choice reasoning (300 samples).

🌟 Overview

KoarAI/LFM2.5-350M-Thinking-0003 is an ultra-compact, high-efficiency hybrid reasoning language model featuring native Chain-of-Thought (CoT) thinking capabilities.

Built upon the state-of-the-art Liquid Foundation Model architecture (LiquidAI/LFM2.5-350M), this model underwent 100% Full Parameter Fine-Tuning in bfloat16 precision. Despite having only 350 Million parameters, it produces structured internal step-by-step logic inside native <think> ... </think> blocks before delivering clean, concise final answers.

💡 Native Thinking Mode Example

<|im_start|>user
Сколько букв 'р' в слове 'программирование'?<|im_end|>
<|im_start|>assistant
<think>
1. Слово: 'программирование'.
2. Разберем по буквам: п-р-о-г-р-а-м-м-и-р-о-в-а-н-и-е.
3. Позиции буквы 'р':
   - 2-я буква: 'р' (п-Р-о...)
   - 5-я буква: 'р' (...г-Р-а...)
   - 10-я буква: 'р' (...и-Р-о...)
4. Итого: 3 буквы 'р'.
</think>
В слове «программирование» содержится **3** буквы «р».<|im_end|>

⚡ Quickstart (Transformers)

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "KoarAI/LFM2.5-350M-Thinking-0003"

tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16 if torch.cuda.is_available() else torch.float32,
    device_map="auto",
    trust_remote_code=True
)

messages = [
    {"role": "user", "content": "Solve step-by-step: 32 + 32 - 42"}
]

prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

outputs = model.generate(
    **inputs,
    max_new_tokens=512,
    temperature=0.6,
    top_p=0.9,
    do_sample=True,
    pad_token_id=tokenizer.eos_token_id
)

print(tokenizer.decode(outputs[0][inputs['input_ids'].shape[1]:], skip_special_tokens=False))

⚙️ Training Specifications

Parameter	Specification
Base Model	`LiquidAI/LFM2.5-350M`
Fine-Tuning Method	Full Fine-Tuning (100% parameters)
Precision	`bfloat16` / `fp16`
Epochs	2.1
Learning Rate	`2.5e-5` (Cosine schedule with 5% warmup)
Effective Batch Size	8 (`batch_size=2`, `gradient_accumulation_steps=4`)
Max Sequence Length	512 tokens
Special Tokens	`<think>`, `</think>`

📦 GGUF Quantized Versions

Quantized GGUF versions for llama.cpp, Ollama, LM Studio, and Jan are available at:
👉 KoarAI/LFM2.5-350M-Thinking-0003-GGUF

🐨 Maintained by KoarAI Lab

Released for the open-source AI community by KoarAI.

Downloads last month: 1,022

Safetensors

Model size

0.4B params

Tensor type

BF16

Model tree for KoarAI/LFM2.5-350M-Thinking-0003

Base model

LiquidAI/LFM2.5-350M-Base

Finetuned

LiquidAI/LFM2.5-350M

Finetuned

(66)

this model

Quantizations

1 model

Space using KoarAI/LFM2.5-350M-Thinking-0003 1

Collection including KoarAI/LFM2.5-350M-Thinking-0003

0003 🔥

Collection

2 items • Updated 1 day ago