---
license: apache-2.0
language:
- en
base_model: LiquidAI/LFM2.5-1.2B-Thinking
model_type: lfm2
library_name: transformers
tags:
- lfm2
- liquid
- reasoning
- cot
- distillate
datasets:
- FlameF0X/Claude-Sonnet-4.6
---

<div align="center">
    <img src="https://cdn-uploads.huggingface.co/production/uploads/6615494716917dfdc645c44e/B0exWWphEmk8Oixke9YlX.png" alt="Liquid Claude Banner" style="width: 100%; max-width: 100%; height: auto; display: inline-block; margin-bottom: 0.5em; margin-top: 0.5em; reading-order: 20px; border-radius: 20px;"/>
</div>

> [!Important]
> Model trained on Claude Sonnet 4.6 Adaptive Thinking (The claude.ai Claude, not the API Claude). The training data is PRIVATE.

# LFM2.5-1.2B-Distilled-Claude-4.6 (Liquid Claude)
LFM2.5-1.2B-Distilled-Claude-4.6 (Liquid Claude) is a distillation of Claude into LFM2.5-1.2B-Thinking via LoRA.

# Training data info:

**THINK BLOCK PATTERNS** <br>
Agentic think blocks (Action/Observation): 8473 <br>
Pure reasoning think blocks: 3005

**CONSECUTIVE USER MESSAGES** <br>
Total consecutive user-user pairs: 319

**MESSAGE LENGTH STATS (chars)**
|   :)    |msgs |avg |max   |
|---------|-----|----|------|
|System   |2325 |2168|2168  |
|User     |11919|332 |39134 |
|Assistant|11738|4386|264340|

Assistant messages total: 11738 <br>
With agentic tool calls in think: 2151 (18.3%)

Total chars in dataset: 60,487,999<br>
Approx tokens (~4 chars/token): 15,121,999

Conversations with <=2 messages (system+1): 121<br>
Conversations with >5 think blocks in a single assistant msg: 319

## Use model
```python
from transformers import pipeline

pipe = pipeline("text-generation", model="FlameF0X/LFM2.5-1.2B-Distilled-Claude-4.6")
messages = [
    {"role": "system", "content": "You are a helpful assistant."}, # I RECOMMEND TO KEEP THIS FOR STABILITY! But you can change the system.
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)
```

## Sample chat:
![image](https://cdn-uploads.huggingface.co/production/uploads/6615494716917dfdc645c44e/BxB99Bu0cLEBV6d01Utmj.png)
(Ignore the fact that it took 1min to reason, i got a i3-6006u / 12GB as hardware and running the f16 quantization)

# Benchmark

|Model                                    |Average|HellaSwag|MMLU |Piqa |Source|
|-----------------------------------------|-------|---------|-----|-----|------|
|FlameF0X/LFM2.5-1.2B-Distilled-Claude-4.6|46.76  |39.51    |31.99|68.77|[Intel/low bit open llm leaderboard](https://huggingface.co/spaces/Intel/low_bit_open_llm_leaderboard)|
|FlameF0X/LFM2.5-1.2B-Thinking-CodeX      |45.25  |39.70    |26.56|69.48|As the one from above|