Compare AI models

Specifications
Parameters
—16B
Context window
200k32k
Benchmarks
HumanEval
84.9%57.3%
MMLU
88.2%55.7%
GSM8K
95%72%
Overview
CompanyAnthropicDeepSeek
Release dateMar 4 2024May 16 2024
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Claude 3 Opus leads DeepSeek V2 Lite on 3 of the 3 benchmarks they both report (HumanEval, MMLU, GSM8K). Claude 3 Opus shipped 73 days before DeepSeek V2 Lite, so benchmark comparisons should account for the intervening progress.

Context windows are 200k (Claude 3 Opus) vs 32k (DeepSeek V2 Lite). Claude 3 Opus is closed, while DeepSeek V2 Lite is open weight.

On HumanEval, Claude 3 Opus leads at 84.9% vs DeepSeek V2 Lite at 57.3%. On MMLU, Claude 3 Opus leads at 88.2% vs DeepSeek V2 Lite at 55.7%. On GSM8K, Claude 3 Opus leads at 95% vs DeepSeek V2 Lite at 72%.