Liquid AI

company

https://www.liquid.ai

AI & ML interests

A new generation of foundation models from first principles.

Recent Activity

mlabonne new activity about 3 hours ago

LiquidAI/LFM2-24B-A2B:When LiquidAI/LFM2.5-24B-A2B ?

mlabonne authored a paper 5 months ago

LFM2 Technical Report

mlabonne authored a paper 5 months ago

Zero-Overhead Introspection for Adaptive Test-Time Compute

View all activity

Papers

LFM2 Technical Report

View all Papers

in LiquidAI/LFM2-24B-A2B about 3 hours ago

When LiquidAI/LFM2.5-24B-A2B ?

#18 opened 1 day ago by

testamentaddress01

posted an update about 2 months ago

Post

3191

Big update to llm-datasets, my curated list of datasets and tools for post-training LLMs.

> Added many new datasets
> New "thinking" column
> Refreshed recommended tools.

Thanks to everyone who told me they used it for their research at ICLR, you motivated this update!

2 replies

·

authored 2 papers 5 months ago

LFM2 Technical Report

Paper • 2511.23404 • Published Nov 28, 2025 • 61

Zero-Overhead Introspection for Adaptive Test-Time Compute

Paper • 2512.01457 • Published Dec 1, 2025 • 3

posted an update 6 months ago

Post

10437

New family of 1B models just dropped!

> LiquidAI/LFM2.5-1.2B-Base: 10T → 28T tokens
> LiquidAI/LFM2.5-1.2B-Instruct: new large-scale multi-stage RL
> LiquidAI/LFM2.5-1.2B-JP: our most polite model
> LiquidAI/LFM2.5-VL-1.6B: multi-image multilingual
> LiquidAI/LFM2.5-Audio-1.5B: 8x times faster, no quality loss

Super proud of this release 🤗

3 replies

·

authored a paper 6 months ago

LFM2 Technical Report

Paper • 2511.23404 • Published Nov 28, 2025 • 61

authored a paper 7 months ago

LFM2 Technical Report

Paper • 2511.23404 • Published Nov 28, 2025 • 61

fernandofernandes

authored 3 papers 7 months ago

Spectrum: Targeted Training on Signal to Noise Ratio

Paper • 2406.06623 • Published Jun 7, 2024 • 16

Domain Adaptation of Llama3-70B-Instruct through Continual Pre-Training and Model Merging: A Comprehensive Evaluation

Paper • 2406.14971 • Published Jun 21, 2024

Training-Free Tokenizer Transplantation via Orthogonal Matching Pursuit

Paper • 2506.06607 • Published Jun 7, 2025 • 3

authored a paper 7 months ago

LFM2 Technical Report

Paper • 2511.23404 • Published Nov 28, 2025 • 61

fernandofernandes

authored a paper 7 months ago

LFM2 Technical Report

Paper • 2511.23404 • Published Nov 28, 2025 • 61

authored a paper 7 months ago

MultiBanana: A Challenging Benchmark for Multi-Reference Text-to-Image Generation

Paper • 2511.22989 • Published Nov 28, 2025 • 17

authored 3 papers 8 months ago

EQ-Bench: An Emotional Intelligence Benchmark for Large Language Models

Paper • 2312.06281 • Published Dec 11, 2023 • 2

Democratizing Diplomacy: A Harness for Evaluating Any Large Language Model on Full-Press Diplomacy

Paper • 2508.07485 • Published Aug 10, 2025 • 10

Antislop: A Comprehensive Framework for Identifying and Eliminating Repetitive Patterns in Language Models

Paper • 2510.15061 • Published Oct 16, 2025 • 3

authored a paper 9 months ago

Luth: Efficient French Specialization for Small Language Models and Cross-Lingual Transfer

Paper • 2510.05846 • Published Oct 7, 2025 • 3

posted an update 9 months ago

Post

8477

LiquidAI/LFM2-8B-A1B just dropped!

8.3B params with only 1.5B active/token 🚀

> Quality ≈ 3–4B dense, yet faster than Qwen3-1.7B
> MoE designed to run on phones/laptops (llama.cpp / vLLM)
> Pre-trained on 12T tokens → strong math/code/IF

1 reply

·

authored 2 papers 9 months ago

Detecting Corpus-Level Knowledge Inconsistencies in Wikipedia with Large Language Models

Paper • 2509.23233 • Published Sep 27, 2025 • 4

CHURRO: Making History Readable with an Open-Weight Large Vision-Language Model for High-Accuracy, Low-Cost Historical Text Recognition

Paper • 2509.19768 • Published Sep 24, 2025 • 7