Skip to main content

OpenAI · Text

GPT-OSS models: Versions, specs, GPU requirements & hosting

OpenAI GPT-OSS provides two primary open-weight MoE checkpoints, 120B and 20B, as sibling deployment sizes.

Current releases Browse engineering services Compare models
Technical deployment reference illustration for GPT-OSS
Reference deployment flow; exact architecture facts are listed separately.

Overview and history

GPT-OSS is an official OpenAI model family represented here through independently verified release, checkpoint, license, runtime and hardware evidence. OpenAI GPT-OSS provides two primary open-weight MoE checkpoints, 120B and 20B, as sibling deployment sizes. Current public releases are grouped by the General tracks so sibling sizes and specialist branches are not presented as false chronological generations.

FusionPointAI currently tracks 9 certified public GPT-OSS releases. The catalogue preserves release dates and evaluates Current, Recent, Previous and Legacy status within each model track rather than by database insertion order.

Two deployment scales
Native MXFP4 checkpoints
Reasoning-oriented open weights

Architecture and model range

The verified public range spans 21B to 117B total parameters and context windows up to 131,072 tokens. Each release states whether it is dense or mixture-of-experts and records active parameters separately when the official source publishes them.

Public releases9
Model tracks1
Current generationGPT-OSS 120B
Last verified2026-09-03T00:00:00.000000Z

Generations, tracks and variants

General

gpt-oss-120b-eagle3-long-context

Recent generation
gpt-oss-120b-Eagle3-long-context

Not published by manufacturer parameters · N/A / Not directly comparable context

Fine-tune Unknown / Not Recommendable

gpt-oss-120b-eagle3-throughput

Recent generation
gpt-oss-120b-Eagle3-throughput

Not published by manufacturer parameters · N/A / Not directly comparable context

Fine-tune Unknown / Not Recommendable

gpt-oss-120b-eagle3-short-context

Recent generation
gpt-oss-120b-Eagle3-short-context

Not published by manufacturer parameters · N/A / Not directly comparable context

Fine-tune Unknown / Not Recommendable

granitelib-rag-gpt-oss-r1-0

Recent generation
granitelib-rag-gpt-oss-r1.0

Not published by manufacturer parameters · N/A / Not directly comparable context

Fine-tune

gpt-oss-120b-eagle3-v3

Recent generation
gpt-oss-120b-Eagle3-v3

Not published by manufacturer parameters · N/A / Not directly comparable context

Fine-tune Unknown / Not Recommendable

gpt-oss

Current generation 4 sibling variants
gpt-oss-safeguard-120b

Not published by manufacturer parameters · N/A / Not directly comparable context

Fine-tune
GPT-OSS 20B

21.0B parameters · 131,072 context

MoE GptOssForCausalLM
gpt-oss-safeguard-20b

Not published by manufacturer parameters · N/A / Not directly comparable context

Fine-tune
GPT-OSS 120B

117.0B parameters · 131,072 context

MoE GptOssForCausalLM

Deployment ecosystem

Choose a specific GPT-OSS release and certified checkpoint before sizing hardware. FusionPointAI rejects unknown runtime configurations, calculates resident memory from total model weights, then distinguishes globally compatible GPUs from currently rentable inventory.

Validate the exact checkpoint license, runtime version, precision kernel, context length, concurrency and multi-GPU topology. Active MoE parameters affect compute, but resident model memory continues to use total stored weights.

Common use cases

Private reasoning
Developer assistants
Local experimentation

License overview

Licenses are certified per release, not inferred from the provider. Review the release page before commercial deployment because terms can differ between generations and variants.

Commercial-use status is a source-backed catalogue classification, not legal advice. Follow the linked official license and attribution requirements.

Release timeline

openai/gpt-oss/gpt-oss-safeguard-120b @ 7c72ded49d1153c13d9ca03792b912bff490f4fbc95bb3f3e866627364e877d0

Official source

openai/gpt-oss/gpt-oss-120b-eagle3-long-context @ 7cf1d00314140a27295699b436a6ab51124fd344b2488461db0d119527407163

Official source

openai/gpt-oss/gpt-oss-safeguard-20b @ 6ba32890463e3ebf15c8a9f9899709094f8185674820fc3ed539846f146a6b0d

Official source

openai/gpt-oss/gpt-oss-120b-eagle3-short-context @ ea3074be24aab39e2e0e4682794bc91bd99755dd871c24b782c99fd3ae06f293

Official source

openai/gpt-oss/gpt-oss-120b-eagle3-throughput @ cd240616aaeae271f313a906b6cb9466a58cf43e5f14d2baf2106a9bebc3c60a

Official source

openai/gpt-oss/gpt-oss-120b-eagle3-v3 @ f68f09dfb1204e915c0b8a74092dbb7324657da69d91b2f0fb9cfd70326d3048

Official source

openai/gpt-oss/granitelib-rag-gpt-oss-r10 @ 8f80f62d210eec39d6b61a1f3d5a3f423510cafbef406069b8a9111ec03ec61d

Official source

openai/gpt-oss/gpt-oss-120b @ 776a162c0c69e2eae8df64b37ac803df6f2dfe928aad74f49583f0532c30e946

Official source

openai/gpt-oss/gpt-oss-20b @ 0092e0d9951eb4f12c2a37d27e03f6c2688cb749e08312cd3145232e3b737168

Official source

Frequently asked questions

Which GPT-OSS release is current?

Current releases are GPT-OSS 120B, GPT-OSS 20B, evaluated within their respective tracks.

Can GPT-OSS be self-hosted?

Only releases with a verified official checkpoint and supported runtime configuration are presented as deployable.

Which GPU should run GPT-OSS?

GPU selection depends on the exact variant, checkpoint precision, context length and concurrency; use the release hardware recommendations rather than the family name alone.

Official references

Last verified: 2026-09-03