LiquidAI: LFM2-8B-A1B

LiquidAI: LFM2-8B-A1B

liquid · Released Oct 20, 2025 Emerging
Intelligence #300 / 583
29.5 Our Score
AA Index #367 / 380
1.8 Artificial Analysis
Input #127 / 583
$0.010 per 1M tokens
Output #127 / 583
$0.020 per 1M tokens
Context #400 / 583
32,768 tokens

Analysis Summary

LFM2-8B-A1B is a compact text model from LiquidAI, built on their liquid foundation model architecture. With an intelligence index of 1.8 and a coding index of 2.3, it sits at the lower end of the capability spectrum, and it lacks vision, tool use, or function calling.

For businesses, practical use cases are narrow: simple text generation, basic summarisation, or as a low-cost pre-processing step in a larger pipeline. The 32K context window is a meaningful constraint for document-length tasks. Its agentic score of 10.5 is low, and there is no evidence of reliable multi-step task execution.

At $0.01 input and $0.02 output per million tokens, it is cheap. It may be worth evaluating for very high-volume, low-stakes text tasks where cost is the primary driver, but businesses should not rely on it for anything requiring consistent reasoning or structured output quality.

Assessed June 17, 2026

Editorial notes

LiquidAI LFM2-8B-A1B is a very small, text-only model with minimal reasoning and coding capability, suited only to the simplest tasks at near-zero cost.

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?

Performance Profile

Intelligence1.1Technical0.9Value7.8Content2.5
Intelligence 1.1/10
Technical 0.9/10
Content 2.5/10
Value 7.8/10

How LiquidAI: LFM2-8B-A1B compares

LiquidAI: LFM2-8B-A1B ranks #367 of 380 AI models we track for overall intelligence, #257 of 292 for agentic tasks. Its 33K-token context window is larger than 31% of the models we list. At $0.01 per million input tokens it is cheaper than 78% of comparable models.

About LiquidAI: LFM2-8B-A1B

LFM2-8B-A1B is an efficient on-device Mixture-of-Experts (MoE) model from Liquid AI’s LFM2 family, built for fast, high-quality inference on edge hardware. It uses 8.3B total parameters with only ~1.5B active per token, delivering strong performance while keeping compute and memory usage low—making it ideal for phones, tablets, and laptops.

8B Parameters

Performance Indices

Source: Artificial Analysis

1.8 Intelligence Index
10.5 Agentic Index
25.3 Math Index

Benchmark Scores

Intelligence

GPQA Diamond 34.4% Graduate-level scientific reasoning
HLE 4.9% Humanity's Last Exam
MMLU Pro 50.5% Multi-task language understanding
AIME 2025 25.3% Competition mathematics (2025)
SciCode 6.8% Scientific computing

Technical

LiveCodeBench 15.1% Live coding evaluation
τ²-Bench 10.5% Conversational agent benchmark

Content

IFBench 26.3% Instruction following

Benchmark data from Artificial Analysis and Hugging Face

How does LiquidAI: LFM2-8B-A1B stack up?

Compare side-by-side with other emerging models.

Compare Models

Model Information

OpenRouter ID liquid/lfm2-8b-a1b
Providerliquid
Release Date October 20, 2025
Context Length32,768 tokens
Status Active

Pricing

Token Type Cost per 1M tokens Cost per 1K tokens
Input $0.01 $0.000010
Output $0.02 $0.000020

Frequently asked questions about LiquidAI: LFM2-8B-A1B

How much does LiquidAI: LFM2-8B-A1B cost?

LiquidAI: LFM2-8B-A1B costs $0.01 per million input tokens and $0.02 per million output tokens.

What is the context window of LiquidAI: LFM2-8B-A1B?

LiquidAI: LFM2-8B-A1B has a context window of 32,768 tokens (33K).

Who created LiquidAI: LFM2-8B-A1B?

LiquidAI: LFM2-8B-A1B is developed by LiquidAI and was released on October 20, 2025.