An open-weight GPT reasoning model for self-hosted agents and controllable deployments.
Technical Specifications
- Provider
- OpenAI
- Category
- open-source
- Release Date
- 2025-08
- Batch API Discount
- 50% off async batch
- Consumer Subscription
- Free / Self-Hosted
- Reasoning Effort Control
- Configurable (Low / Med / High)
- Structured Outputs
- Yes
- Output Modalities
- text
- Official Documentation
- Visit Official Docs ↗
Data Privacy, Licensing & Deployment Governance
Official Provider Release Intelligence, API Quotas & Agentic Specs
- Official API Model ID (SDK String)
- gpt-oss-20b
- Supported SDK & API Protocol
- OpenAI Responses API & Chat Completions v1
- Lifecycle & Deprecation Status
- Active GA (Guaranteed support through 2027+)
- Production Rate Limits (RPM & TPM)
- Unlimited (Hardware-bound on self-hosted GPU) / 10,000 RPM Cloud · Unlimited local throughput / 10M+ TPM Managed Cloud
- Long-Context Pricing Tier (>128K / >200K)
- Flat rate ($0.1/1M in) across full 131K window
- Prompt Cache Write & TTL Policy
- Automatic prefix caching at $0.01/1M (no extra cache-write fee)
- Computer Use & Browser GUI Agent
- API / CLI Tool-Use & DOM Agent ready (no raw pixel desktop click loop)
- MCP & Parallel Tool Calling
- Native MCP Client/Server ready · Parallel function calling · 100% Strict JSON Schema
- Web Search Grounding & Code Sandbox
- Native Web Search tool + Hosted Code Interpreter & File Search RAG
- Fine-Tuning & Distillation Support
- Full Weights SFT, LoRA, QLoRA, DPO & GRPO RL (Unsloth / Axolotl / Llama-Factory)
- Multimodal Tokenization Billing
- Text BPE (~4 chars/token)
- Weights Precision & Quantization
- BF16 / FP16, FP8 Dynamic, INT4 AWQ / GPTQ, GGUF (Q4_K_M / Q8_0)
- Minimum GPU VRAM Requirement
- 13 GB VRAM (4-bit Quantized) · 45 GB VRAM (Native BF16)
- Supported Inference Engines
- vLLM, SGLang, Ollama, llama.cpp, TensorRT-LLM, HuggingFace TGI
- System Card & Safety Framework
- OpenAI Preparedness Framework Audited & System Card Published
- Enterprise Compliance (SOC2 / HIPAA / ZDR)
- 100% Air-Gapped On-Premises / Private VPC · Full HIPAA, GDPR & SOC2 Control
GPT-OSS 20B Cost, Context, Hardware & API Integration
Pay-as-you-go API is cheaper for this volume (saves $9.20/mo vs $20 flat plan).
Break-even point: ~926 requests/day equals $20/month.
Same Monthly Workload Cost Across Popular Reference Models
Live Prompt Paste Tokenizer & Cost Counter
Paste real system instructions, JSON schemas, or code to measure BPE token counts across tokenizers.
Hybrid Prompt Routing Cascade Simulator
Route routine queries to a fast lightweight model ($0.15/1M) and escalate only complex reasoning turns to GPT-OSS 20B:
Benchmark Scores & 6-Axis Capability Radar
- SWE-bench Verified
- 32.5%
- GPQA Diamond
- 55%
- MMLU-Pro
- 72.5%
- HumanEval
- 84.5%
- Humanity's Last Exam
- 7.2%
- Terminal-Bench
- 22%
- SciCode
- 24.5%
- LiveCodeBench
- 46%
- AIME
- 72%
- Long-Context Reasoning
- 63%
- Chatbot Arena Elo
- 1288pts
Strengths
- Open weights — run it fully on your own hardware
- Small enough for a laptop or single consumer GPU
- Cheap on hosted providers and supports tool calling
- Runs comfortably on consumer hardware
- Apache-style open licence for commercial use
Limitations & Trade-offs
- Clearly weaker than frontier models on hard reasoning
- Text-only — no vision or audio input
- Needs careful prompt tuning to reach its best output
Editorial Assessment
The best open-weight option when you need to own the deployment.
Best for: Self-hosted agents, on-premise deployments and privacy-sensitive work
Compare GPT-OSS 20B
Frequently Asked Questions about GPT-OSS 20B
How much does the GPT-OSS 20B API cost per 1M tokens?+
GPT-OSS 20B by OpenAI is priced at $0.1 per 1M input tokens and $0.5 per 1M output tokens. With prompt caching enabled, cached input tokens cost $0.01 per 1M tokens, and asynchronous Batch API requests receive a 50% discount.
Should I use a $20/month subscription or the GPT-OSS 20B pay-as-you-go API?+
A flat $20/month consumer subscription is predictable for daily interactive chat and manual coding. However, if you are building software, running automated workflows, or sending low-to-moderate volume with prompt caching enabled, the pay-as-you-go API is typically far more cost-effective and offers higher rate limits without manual UI caps.
How many words and pages fit in GPT-OSS 20B's context window?+
GPT-OSS 20B features a 131K tokens context window, which equals approximately 98,304 English words or ~262 standard single-spaced pages in a single prompt. Its maximum output generation per turn is 32,768 tokens.
Can I run GPT-OSS 20B locally on my own GPU (Open Weights)?+
Yes, GPT-OSS 20B (21B (3.6B active MoE)) releases open weights that can be self-hosted locally or on private cloud GPUs using vLLM, Ollama, SGLang, or llama.cpp.
Does OpenAI train on data sent to the GPT-OSS 20B API, and is commercial use allowed?+
100% Isolated when self-hosted on private GPUs or VPC. License tier: Open-Weights Research & Commercial License. Full commercial ownership of generated outputs. Supported deployment channels include: Self-Hosted GPU (vLLM / Ollama), Private VPC, or Managed Serverless API.
How does GPT-OSS 20B perform on coding and reasoning benchmarks?+
GPT-OSS 20B achieves a composite Intelligence Index of 46/100, scoring 32.5% on SWE-bench Verified, 55% on GPQA Diamond, 72.5% on MMLU-Pro, and 1288 pts on Chatbot Arena Elo.
What is GPT-OSS 20B best used for?+
GPT-OSS 20B is best suited for: Self-hosted agents, on-premise deployments and privacy-sensitive work. Assessment: The best open-weight option when you need to own the deployment.
OpenAI Family Lineage, Routing Tiers & Upgrade Path
Step down to a lighter sibling for high-volume routing or step up to OpenAI's flagship reasoning tier:
GPT-6 Astra
GPT-6.1 Sol
GPT-6 Sol
GPT-6 Luna
GPT-5.6 Sol
GPT-5.6 Terra
AI Glossary & Architecture Concepts
View full glossary →Key machine learning concepts and benchmarks relevant to GPT-OSS 20B:
The basic unit of text that a language model processes, typically representing about 3 to 4 English characters or roughly 0.75 words. For example, the sentence "Artificial intelligence is fast" breaks down into about 5 tokens.
A standardized evaluation suite used to objectively measure and rank an AI model's capabilities across specific tasks — such as SWE-bench Verified for software engineering, AIME for competitive math, or MMLU for general knowledge.
The maximum number of tokens a model can consider at once — both the input it reads and the output it writes. A 1,000,000-token context window can hold roughly 750,000 English words in a single request.
An AI model trained via reinforcement learning to spend dynamic compute thinking through complex logic, math, and architectural problems before outputting a solution. Examples include OpenAI o1, o3, and DeepSeek R1.
A step-by-step reasoning technique where an AI model generates intermediate problem-solving thoughts before arriving at a final answer. Native reasoning models like OpenAI o3 and DeepSeek R1 use CoT to solve complex math and coding challenges.
The ability of a language model to recognize when an external capability is needed and output structured JSON arguments to invoke external APIs, databases, or terminal commands. The runtime executes the tool and passes results back to the model.
Benchmark scores, token prices, and rate limits evolve rapidly as AI labs release model updates. Always verify current specifications and pricing on the official provider documentation.
Sources: OpenAI open-weight model card (Hugging Face) & official pricing↗