Beta version: *Information might not be fully accurate. Please report any discrepancies.

MetaVerifiedOpen Weights3 benchmarks

Llama 3.3 70B

Released 2024-12-0670B Architecture

Training: 2023-12

Latest Data

Unknown

Context Window

128k

tokens

Input Cost

$0.20

per 1M tokens

Output Cost

$0.20

per 1M tokens

Parameters

70B

model footprint

Benchmark Provenance

Performance Analysis // Verified Benchmarks

88.6/ 100

Verified

Last Verified: Unknown DateMeta AI

Massive Multitask Language Understanding covers 57 subjects across STEM, the humanities, social sciences, and more.

73/ 100

Verified

Last Verified: Unknown DateMeta AI

Challenging competition mathematics problems (AIME/IMO level).

89/ 100

Verified

Last Verified: Unknown DateMeta AI

Functional correctness of synthesized programs from docstrings.