Which is better, LFM2.5-1.2B-Instruct or Llama 3.1 Instruct 405B?

Refresh Models —Deals —Regulatory —Sources —

LFM2.5-1.2B-Instruct vs Llama 3.1 Instruct 405B

22 SHARED BENCHMARKS

Across 22 shared benchmarks, LFM2.5-1.2B-Instruct scores higher on 2 and Llama 3.1 Instruct 405B on 18, with 2 level. The widest gap is AA Intelligence, where Llama 3.1 Instruct 405B scores 8.5 against 2.3.

LIQUID AIVSMETA22 SHARED2–18 HEAD-TO-HEAD

BenchmarkLFM2.5-1.2B-InstructMARGINLlama 3.1 Instruct 405B

AA Agentic Index3.66.3

AA Intelligence2.38.5

AA-LCR024.7

AA-Omniscience-72.1-17.1

Artificial Analysis Coding Index0.814.5

critpt00

gdpval00

GPQA Diamond38.951.5

GSM8K64.596.8

HLE6.74.2

humaneval5.389

IFBench47.339

ifeval86.288.6

math_500_em63.273.8

MBPP5.573.4

MMLU-Pro44.473.4

mt_bench3.99.1

OmniScience Accuracy723.2

OmniScience Non-Hallucination14.949

scicode2.329.9

Terminal-Bench Hard06.8

τ²-Bench Telecom (AA run)10.819

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.