Which is better, LFM2.5-1.2B-Instruct or Llama 3.2 Instruct 1B?

Refresh Models —Deals —Regulatory —Sources —

LFM2.5-1.2B-Instruct vs Llama 3.2 Instruct 1B

29 SHARED BENCHMARKS

Across 29 shared benchmarks, LFM2.5-1.2B-Instruct scores higher on 23 and Llama 3.2 Instruct 1B on 4, with 2 level. The widest gap is J-MATH500, where LFM2.5-1.2B-Instruct scores 50 against 11.4.

LIQUID AIVSMETA29 SHARED23–4 HEAD-TO-HEAD

BenchmarkLFM2.5-1.2B-InstructMARGINLlama 3.2 Instruct 1B

AA Agentic Index3.60

AA Intelligence2.31.1

AA-LCR06

AA-Omniscience-72.1-54.7

AIME25 no tools140.3

Artificial Analysis Coding Index0.80.6

Avg.50.118.3

BFCL v349.121.4

critpt00

Domain Avg39.721.2

GPQA Diamond38.928.8

GSM8K64.539

HLE6.75.5

IFBench47.322.8

ifeval86.252.4

J-BFCLv346.321.1

J-GSM8K50.225.2

J-MATH5005011.4

JGPQA31.732.3

JHumanEval+28.717.7

JMMLU47.634

JMMLU-ProX31.415.9

math_500_em63.223.4

MMLU-Pro44.420.8

OmniScience Accuracy77

OmniScience Non-Hallucination14.933.7

scicode2.31.7

Terminal-Bench Hard00

τ²-Bench Telecom (AA run)10.80

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.