LFM2-8B-A1B
LFM2-8B-A1B.
17
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Liquid AI Oct 2025
LFM2-8B-A1B
Open weights Mid-size
111.1
#374 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
- IFBench 25.9% #76 /80
- LiveCodeBench v5 21.4% #47 /47
- MMLU-Pro 37.4% #163 /167
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MGSM | 72.4% | #42 of 57 | Lab model card | |
| IFEval | 77.6% | #113 of 135 | Lab model card | |
| MMLU | 64.8% | #112 of 131 | Lab model card | |
| Multi-IF | 58.2% | #24 of 28 | Lab model card | |
| Wer wird Millionär? LLM Benchmark | 88.0 | #70 of 77 | Official leaderboard | |
| MMMLU | 55.3% | #65 of 71 | Lab model card | |
| IFBench | 25.9% | #76 of 80 | Lab model card | |
| MMLU-Pro | 37.4% | #163 of 167 | Lab model card |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GSM-Plus | 64.8% | #1 of 2 | Lab model card | |
| MATH Level 5 | 62.4% | #1 of 2 | Lab model card | |
| GSM8K | 84.4% | #58 of 92 | Lab model card | |
| MATH-500 | 74.2% | #64 of 70 | Lab model card | |
| GPQA Diamond | 29.3% | #296 of 302 | Lab model card |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HumanEval+ | 69.5% | #23 of 27 | Lab model card | |
| LiveCodeBench v6 | 21% | #90 of 90 | Lab model card | |
| LiveCodeBench v5 | 21.4% | #47 of 47 | Lab model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SnakeBench | 15 | #170 of 188 | Official leaderboard |
Its lab's numbers put it at 111.1, number 374 of 404
The Self-Reported Index fits one score to the 13 capability results Liquid AI reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 108.8 to 113.5. How we compute it.
- GSM8K 84.4
- IFEval 77.6
- MATH-500 74.2
- MGSM 72.4
- HumanEval+ 69.5
- MMLU 64.8
- Multi-IF 58.2
- MMMLU 55.3
- MMLU-Pro 37.4
- GPQA Diamond 29.3
- IFBench 25.9
- LiveCodeBench v5 21.4
- LiveCodeBench v6 21
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.
Its system card
Every result in the longer documents the lab published with it.
LFM2 Technical Report Nov 28, 2025