gpt-oss-20b
Official · MXFP4 · safetensors
Runtime certified: Transformers
| Precision / quantization | Format | Bits | Publisher trust | Deployment state |
|---|---|---|---|---|
| MXFP4 | safetensors | 4 | Official | Runtime certified |
GPT-OSS · General · GPT-OSS
GPT-OSS 20B: Specs, GPU requirements, hosting & deployment
GPT-OSS 20B is a current General release in the GPT-OSS family, certified from its immutable official source revision.
Total parameters21.0B
Active parameters3.6B
Context window131,072 tokens
Release statusCurrent · Current
ArchitectureGptOssForCausalLM
ModalitiesText
Licenseapache-2.0
Deployment stateDeployable with certified configuration
Commercial stateCommercial Use Allowed
Release date2025-08-04
Architecture, strengths and use cases
GPT-OSS 20B is an official self-hostable OpenAI release in the GPT-OSS General track. It uses GptOssForCausalLM, carries 21B / 3.6B MoE, supports a verified context window of 131,072 tokens, and exposes only source-certified checkpoints and runtime combinations. Its release, license and deployment facts are tied to the official repository revision rather than inferred from the family name.
GPT-OSS 20B uses GptOssForCausalLM with 21B / 3.6B MoE. The stored architecture profile records 24 layers, hidden width 2880, 64 attention heads and a 131,072 token context ceiling where the official configuration publishes those fields.
Verified GptOssForCausalLM architecture
131,072 token certified context
Official MXFP4, Unknown checkpoint coverage
Private LLM inference
Agentic workflows
Long-context document processing
Certified checkpoints and quantizations
Only exact, source-certified variants attached to this release are shown.
Recommended hardware
Use only the listed transformers runtime configuration with the exact checkpoint precision. The recommendation engine checks weight memory, KV cache, runtime overhead, safety headroom, GPU architecture and topology before ranking rentable and reference hardware.
Estimated weight memory10.07 GB
KV cache0.09 GB
Runtime overhead6.55 GB
Total with safety margin20.73 GB
Recommended best value
1× NVIDIA H100 SXM 80GB
80 GB total VRAM · nvswitch
59.27 GB headroom · Available
USD 1,860 / monthLowest-cost viable
1× NVIDIA H100 PCIe 80GB
80 GB total VRAM · pcie
59.27 GB headroom · Limited capacity
USD 1,579 / monthHighest performance
1× NVIDIA H200 SXM 141GB
141 GB total VRAM · nvswitch
120.27 GB headroom · Limited capacity
USD 2,345 / monthLicense and self-hosting considerations
The full 21B resident weight set must fit across the selected GPUs. Context and concurrency add KV-cache memory; MoE active parameters do not replace resident-weight memory.
GPT-OSS 20B is published under apache-2.0. The release-scoped official license link and verification date are retained with the catalogue record.
Commercial Use Allowed. The official license permits commercial use; its obligations still apply.
Alternatives and sibling variants
GPT-OSS 120B, gpt-oss-120b-Eagle3-v3, granitelib-rag-gpt-oss-r1.0
Release timeline
openai/gpt-oss/gpt-oss-20b
Source details
Source revision: 0092e0d9951eb4f12c2a37d27e03f6c2688cb749e08312cd3145232e3b737168
Frequently asked questions
How much VRAM does GPT-OSS 20B require?
The exact estimate depends on checkpoint precision, context and concurrency; the page recommendation uses the certified memory engine and compatible runtime configurations.
Which runtimes support GPT-OSS 20B?
Certified upstream runtime evidence is available for transformers.
Can GPT-OSS 20B be used commercially?
The stored commercial classification is Commercial Use Allowed. Consult the official apache-2.0 terms for the final obligations.
Official references
Last verified: 2026-09-03