gpt-oss-20b

OpenAI
Model overview

gpt-oss-20b

Text-only open-weight reasoning model for low-latency, local, and edge deployment, designed to run within 16GB of memory.

general LLMreasoningcodingagentlocal inferencecode generationsoftware engineering agent
Representative version

gpt-oss-20b

CompanyOpenAI
Release2025-08-05
Release typeModel release
Parameters21B / 3.6B active
ArchitectureMixture-of-experts transformer
Context131,072
Input modalitiestext
Output modalitiestext
Tool use / function callingConfirmed: Tool use, Function calling
Structured output / reasoningConfirmed: Structured output, Reasoning
Vision / audio / videoNot shown: Vision, Audio, Video
Open weights / LicenseWeights available / Apache-2.0 (license source pending verification)
openai-gpt-oss-20b-current · License source pending verification
Vendor APINot hosted on OpenAI API or ChatGPT
gpt-oss Model Card | OpenAIopenai-gpt-oss-20b-current · 2025-08-05
OpenRouterOpenRouter listed
PricingOpenRouter · 2026-07-15: $0.03 / $0.13
input / output, per 1M tokens
BenchmarkNo reliable public source
Public signals1 activity snapshot
Use caseslocal reasoning / low-latency coding / edge deployment / local inference
Claim evidenceRelease eventgpt-oss Model Card | OpenAIgpt-oss-20b · 2025-08-05
Model specs / weightsIntroducing gpt-oss | OpenAIgpt-oss-20b · gpt-oss-20b
OpenRouter

Pricing / context / availability

OpenRouterOpenRouter listed
OpenRouter listingopenai/gpt-oss-20b
Mapped versiongpt-oss-20b
openai-gpt-oss-20b-current
PricingOpenRouter · 2026-07-15: $0.03 / $0.13
input / output, per 1M tokens
Context131072
Speed/LatencyNo reliable public source
Snapshot time2026-07-15T20:11:41.551Z
SourceView source
Public activity counters

Cumulative Hugging Face model-card downloads / likes / spaces

PlatformEntryDownloadsLikesSpacesTimeSource
huggingfaceopenai/gpt-oss-20b7,294,1714,7961002026-07-15T20:11:41.551ZView source
Evidence gaps

Public display boundary

No reliable benchmark sourceNo linkable real reviewSpeed/latency not disclosed