Which is better, Gemma 3 1B Instruct or LFM2.5-1.2B-Instruct?

Refresh Models —Deals —Regulatory —Sources —

Gemma 3 1B Instruct vs LFM2.5-1.2B-Instruct

31 SHARED BENCHMARKS

Across 31 shared benchmarks, Gemma 3 1B Instruct scores higher on 3 and LFM2.5-1.2B-Instruct on 25, with 3 level. The widest gap is humaneval, where Gemma 3 1B Instruct scores 41.5 against 5.3.

GOOGLEVSLIQUID AI31 SHARED3–25 HEAD-TO-HEAD

BenchmarkGemma 3 1B InstructMARGINLFM2.5-1.2B-Instruct

AA Agentic Index3.53.6

AA Intelligence12.3

AA-LCR00

AA-Omniscience-75.5-72.1

AIME25 no tools114

Artificial Analysis Coding Index0.20.8

Avg.24.650.1

BFCL v316.649.1

critpt00

Domain Avg23.939.7

GPQA Diamond24.238.9

GSM8K62.864.5

HLE5.36.7

humaneval41.55.3

IFBench20.547.3

ifeval80.286.2

J-BFCLv317.346.3

J-GSM8K33.650.2

J-MATH50015.650

JGPQA24.231.7

JHumanEval+2528.7

JMMLU34.547.6

JMMLU-ProX14.131.4

math_500_em45.263.2

MBPP35.25.5

MMLU-Pro14.744.4

OmniScience Accuracy3.87

OmniScience Non-Hallucination17.714.9

scicode0.72.3

Terminal-Bench Hard00

τ²-Bench Telecom (AA run)10.510.8

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.