The Open-Source AI Stack
RSS
All models

Models · Compare

Llama 3.1 405B Instruct vs Claude 3.5 Sonnet

Rows highlighted in warm gray are where the models differ. Numbers carry their as-of date and primary source.

Specs

Field A: Llama 3.1 405B Instruct B: Claude 3.5 Sonnet
Released 2024-07-232024-06-20
Developer MetaAnthropic
Openness Source-availableProprietary
License Llama 3.1 Community LicenseProprietary
OSI-approved nono
Data released nono
Training code nono
Architecture denseunknown
Total params 405B
Active params
Experts
Context window 131K200K
Attention gqaunknown
Position enc. rope-llama3unknown
Pretraining tokens
Post-training sft, dpo, rejection-samplingrlhf, constitutional
Training hardware H100
$/M input
$/M output
Output tok/sec

Benchmarks

Missing scores render as not reported; never inferred. Bold highlights the leader per benchmark.

General reasoning

Code

Context · A

The first open-weights model trained at GPT-4-comparable scale. Pretraining used 16K H100 GPUs and a published tech report; the release reset expectations for how open the frontier could be.

Context · B

The first Claude release to beat its own larger sibling (Claude 3 Opus) on most benchmarks. Established Artifacts, driving a wave of code-and-canvas product copies.

Llama 3.1 405B Instruct detail → · Claude 3.5 Sonnet detail →