NOUMENOpen app
NOUMEN

OpenAI · Open weights

GPT OSS 20B

Added Aug 5, 2025Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

An open-weight 21B parameter model released under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for lower-latency inference and deployability on consumer or single-GPU hardware. The model is trained in OpenAI's Harmony response format and supports reasoning level configuration, fine-tuning, and agentic capabilities including function calling, tool use, and structured outputs.

  • Reasoning
  • Tools
Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.02
Output
$0.10

Reference prices as published by the model provider. API example

Specifications

Context window
131.1K
Max output
16.4K
Average output (7 days)
1.9K tokens
Longer than 91% of 468 models

Side by side

GPT OSS 20B compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

GPT OSS 20BOpenAIThis modelGPT-OSS 120B TEEOpenAIRemove
ProviderOpenAIOpenAI
Discovery tierOpen weightsOpen weights
AddedAug 5, 2025Jan 1, 2024
Limits
Context window131.07K131.07K
Max output16.38K16.38K
Average output1.9K tokens174 tokens
Pricing, USD per 1M tokens
Input$0.02Best$2.00
Output$0.10Best$2.00
Cache read—$2.00
Performance
Request success (24h)99.6%—
Benchmark indexes
Intelligence Index9.011.6Best
Coding Index20.730.4Best
Math Index89.393.4Best
Capabilities
ReasoningYesNo
ToolsYesNo
Reasoning effort——

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • GPT OSS 20B
  • GPT-OSS 120B TEE

Intelligence Index

  • GPT OSS 20B9.0
  • GPT-OSS 120B TEE11.6

GPT OSS 20B tested at high reasoning effort

GPT OSS 20B: Better than 16% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • GPT OSS 20B: 20.7
    • GPT-OSS 120B TEE: 30.4Top

  • Math Index

    Composite of the math evaluations

    • GPT OSS 20B: 89.3
    • GPT-OSS 120B TEE: 93.4Top

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • GPT OSS 20B: 0.2%
    • GPT-OSS 120B TEE: 0.2%Top

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • GPT OSS 20B: 0.0%Top
    • GPT-OSS 120B TEE: 0.0%Top

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • GPT OSS 20B: 339 Elo
    • GPT-OSS 120B TEE: 608 EloTop

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • GPT OSS 20B: 2.0%
    • GPT-OSS 120B TEE: 4.0%Top

  • AA-LCRv1.1

    Reasoning across long documents

    • GPT OSS 20B: 34.7%
    • GPT-OSS 120B TEE: 52.0%Top

  • MLCR-AA

    Long-context reasoning over medical records

    • GPT OSS 20B: 1.1%
    • GPT-OSS 120B TEE: 1.7%Top

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • GPT OSS 20B: 11.0%
    • GPT-OSS 120B TEE: 19.6%Top

  • GPQA Diamond

    Graduate-level science questions

    • GPT OSS 20B: 68.8%
    • GPT-OSS 120B TEE: 78.2%Top

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • GPT OSS 20B: 74.8%
    • GPT-OSS 120B TEE: 80.8%Top

  • IFBench

    Following precise output instructions

    • GPT OSS 20B: 65.1%
    • GPT-OSS 120B TEE: 69.0%Top

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • GPT OSS 20B: 0.0%Top
    • GPT-OSS 120B TEE: 0.0%Top

  • SciCode

    Scientific computing in Python

    • GPT OSS 20B: 38.9%Top
    • GPT-OSS 120B TEE: 34.0%

  • LiveCodeBench

    Recent competitive programming problems

    • GPT OSS 20B: 77.7%
    • GPT-OSS 120B TEE: 87.8%Top

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • GPT OSS 20B: 89.3%
    • GPT-OSS 120B TEE: 93.4%Top

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • GPT OSS 20B: 10.6%
    • GPT-OSS 120B TEE: 23.5%Top

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • GPT OSS 20B: 60.2%
    • GPT-OSS 120B TEE: 65.8%Top

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • GPT OSS 20B: 34.7%
    • GPT-OSS 120B TEE: 52.0%Top

Scores from Artificial Analysis, updated Oct 26, 2026. GPT OSS 20B was tested at high reasoning effort; compare scores measured at the same effort.Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Prompt cache hits
0.0%
Request success (24h)
99.6%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 77.7%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-oss-20b",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="openai/gpt-oss-20b",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

OpenAI

Other OpenAI models, starting with the same family.