Claude 3.5 SonnetvsDeepSeek-V4-Pro

Claude 3.5 Sonnet
DeepSeek-V4-Pro
API pricing
Cheapest input
$0.7078StreamLake
Cheapest output
$1.4157StreamLake
Benchmarks
BullshitBench v2
45%14%
GPQA Diamond
59.4%90.1%
Benchmarks
SWE-Bench Verified
33.4%
LiveCodeBench
93.5%
BrowseComp
83.4%
Overview
CompanyAnthropicDeepSeek
Release dateJun 20 2024Apr 24 2026
AccessProprietaryOpen Weight

Other comparisons

Claude 3.5 SonnetvsGPT-6 AstraDeepSeek-V4-ProvsGPT-6 AstraClaude 3.5 SonnetvsGemini 3.8 FlashDeepSeek-V4-ProvsGemini 3.8 FlashClaude 3.5 SonnetvsMuse Spark 1.3DeepSeek-V4-ProvsMuse Spark 1.3Claude 3.5 SonnetvsGrok 4.6DeepSeek-V4-ProvsGrok 4.6Claude 3.5 SonnetvsMistral Medium 3.5DeepSeek-V4-ProvsMistral Medium 3.5Claude 3.5 SonnetvsKimi K3DeepSeek-V4-ProvsKimi K3

Frequently asked questions

Claude 3.5 Sonnet and DeepSeek-V4-Pro are evenly matched across the 2 benchmarks they both report (BullshitBench v2, GPQA Diamond). Claude 3.5 Sonnet shipped 673 days before DeepSeek-V4-Pro, so benchmark comparisons should account for the intervening progress.

Claude 3.5 Sonnet is proprietary, while DeepSeek-V4-Pro is open weight.

On BullshitBench v2, Claude 3.5 Sonnet leads at 45% vs DeepSeek-V4-Pro at 14%. On GPQA Diamond, DeepSeek-V4-Pro leads at 90.1% vs Claude 3.5 Sonnet at 59.4%.