Llama 3.1 405B

Announced Jul 23, 2024

A larger downloadable flagship

What could it do?

The largest model let developers generate synthetic data, shrink it into smaller models, and fine tune it for their own apps.

What changed?

Expanded the Llama family with a 405B parameter model.

WHY IT MATTERED

Model access

Broadened options for organizations hosting their own models.

Where it fell short

Large hardware needs and license conditions affect use.

Research draft

Capabilities describe the developer’s announcement. This entry has not completed source review.

Read original source

Benchmark results

EPOCH AI CAPABILITY ESTIMATE

128.8index points

Tested variant: Llama 3.1-405BSource interval: 123.8 to 130.8Variant date in source: Jul 23, 2024

A benchmark estimate, not a percentage or capability multiplier. Reasoning settings are not specified in this source table. Historical estimates can change in later snapshots.

Epoch AI methodology ↗Download the source snapshotChecked Oct 7, 2026 · CC BY 4.0

PUBLISHED BENCHMARK RESULT

1335rating points

Tested: llama-3.1-405b-instruct-bf16Text Arena Overall · October 8, 2026Reported interval: 1331 to 133941,375 votes

One explicitly named variant per release. Scores come from the same Overall snapshot; preliminary entries and reported intervals are preserved. These are current ratings of earlier variants, not their launch day ratings.

Source: Text Arena ↗Download selected resultsChecked Oct 8, 2026

Understand the benchmarks Find this release on the chart