Compare AI models

Specifications
Parameters
671B21B
Context window
128k128k
API pricing
Cheapest input
$0.256DeepInfra$0.018AkashML
Cheapest output
$0.712DeepInfra$0.09AkashML
$103StreamLake$8.10Darkbloom · fp8
Benchmarks
SWE-Bench Verified
42%60.7%
GPQA Diamond
59.1%71.5%
MMLU
88.5%85.3%
Overview
CompanyDeepSeekOpenAI
Release dateDec 26 2024Aug 5 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

gpt-oss-20b leads DeepSeek V3 on 2 of the 3 benchmarks they both report (SWE-Bench Verified, GPQA Diamond, MMLU). DeepSeek V3 shipped 222 days before gpt-oss-20b, so benchmark comparisons should account for the intervening progress.

DeepSeek V3 has 671B parameters, while gpt-oss-20b has 21B. Context windows are 128k (DeepSeek V3) vs 128k (gpt-oss-20b).

On SWE-Bench Verified, gpt-oss-20b leads at 60.7% vs DeepSeek V3 at 42%. On GPQA Diamond, gpt-oss-20b leads at 71.5% vs DeepSeek V3 at 59.1%. On MMLU, DeepSeek V3 leads at 88.5% vs gpt-oss-20b at 85.3%.