gpt-oss-20b

gpt-oss-20b · Family: gpt-oss

OpenAI open-weight MoE language model (21B/3.6B active) with native MXFP4 quantization; fits within 16GB of memory.

✓ Active✓ Public access⚖ Open weightsReasoning modelLLMTool-using model📁 gpt-oss

Context window

128K

tokens

Parameters

21B (3,6B aktywnych)

parameters

Max output

131,072

tokens

Release date

5 August 2025

🏢OpenAIProducer

Access:DownloadAPIHostedDeployment:💻 Local☁ Cloud📱 On-device

Overview

gpt-oss-20b is an open-weight language model from OpenAI, released on August 5, 2025 under the Apache 2.0 license. It is a Mixture-of-Experts model with 21B parameters, of which 3.6B are active per token (32 experts, 4 active per token, 24 layers). The model supports a context window of up to 131,072 tokens.

Thanks to native MXFP4 quantization of the MoE-layer weights, the model fits within 16GB of memory, enabling local execution on consumer hardware and edge devices. It offers configurable reasoning effort (low, medium, high), full chain-of-thought access, function calling, web browsing, Python code execution, and structured outputs. The model uses the harmony response format and supports fine-tuning.

Classification

Reasoning modelLLMTool-using model

Family: gpt-oss

Access & deployment

DownloadAPIHosted

LocalCloudOn-device

Weights: Open weights

Key parameters

📏 Context: 128K

🧩 Parameters: 21B (3,6B aktywnych)

✓ Tools · ✓ Fine-tuning

📥 Input: text

Technical specification

Context window

128K

tokens

Parameters

21B (3,6B aktywnych)

parameters

Max output tokens

131,072

tokens per response

Knowledge cutoff

1 Jun 2024

Knowledge boundary

License

Apache 2.0

Hardware requirements

Fits within 16GB of memory thanks to native MXFP4 quantization of the MoE weights; can run locally on consumer hardware and edge devices.

Features:✓ Tool use✓ Fine-tuning

Modalities

⬇ Input

text

⬆ Output

textcode

Capabilities and applications

Native model capabilities

Advanced reasoning

The ability to perform multi-step, structured reasoning: analysing problems, planning steps, and drawing conclusions from hypotheses. Reasoning-first models (e.g. GPT-5.1 Thinking) dedicate a portion of inference to chains of thought before responding.

Category: reasoning

Mathematical reasoning

The model's ability to solve mathematical tasks requiring multi-step reasoning — equations, proofs, combinatorics, geometry, calculus and competition-level problems.

Category: reasoning

Coding

Generating, analysing and modifying code in many programming languages. Covers writing functions, debugging, refactoring, code review, and creating tests. Measured by benchmarks such as HumanEval and SWE-bench.

Category: coding

Agentic capability

The model's ability to autonomously plan and execute multi-step tasks by sequentially using tools, maintaining context, and adapting to intermediate results.

Category: planning

Tool use

The model's ability to call external functions, APIs and tools during a conversation: calculator, search engine, code editor, database. The model decides when and how to use a tool and interprets its result.

Category: planning

Structured output

Producing data in structured formats such as JSON.

Category: structured_generation

Long context

Support for large context windows — tens to hundreds of thousands (or millions) of input tokens. Enables analysis of entire codebases, long documents, and many parallel conversations without losing earlier information. GPT-5.1 supports 400,000 tokens.

Category: language

Adaptive reasoning effort

The model decides how much 'thinking' to allocate to a given query: simple questions are answered quickly, complex problems receive more inference cycles. A GPT-5.1 feature (both Instant and Thinking) that shortens time on easy tasks and extends it for hard ones.

Category: reasoning