NOUMENOpen app
NOUMEN

OpenAI · Open weights

GPT-OSS 120B TEE

Added Jan 1, 2024Open weights

Compare models

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

Loading models…

Use in the app

Open-source GPT model with 120B parameters. running inside a TEE (Trusted Execution Environment), with provider attestation support.

Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$2.00
Output
$2.00
Cache read
$2.00

Reference prices as published by the model provider. API example

Specifications

Context window
131.1K
Max output
16.4K
Average output (7 days)
174 tokens
Longer than 18% of 468 models

Side by side

GPT-OSS 120B TEE 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 120B TEEOpenAIThis modelMistral Small 4 119B ThinkingMistralRemove
ProviderOpenAIMistral
Discovery tierOpen weightsOpen weights
AddedJan 1, 2024Mar 17, 2026
Limits
Context window131.07K262.14KBest
Max output16.38K16.38K
Average output174 tokens997 tokens
Pricing, USD per 1M tokens
Input$2.00$0.40Best
Output$2.00$1.40Best
Cache read$2.00—
Performance
Throughput—155.4 tok/s
Time to first token—4.0 s
Request success (24h)—96.2%
Benchmark indexes
Intelligence Index11.6Best11.3
Coding Index30.4Best26.6
Math Index93.4—
Capabilities
ReasoningNoYes
ToolsNoYes
VisionNoYes
Structured outputNoYes
Reasoning effort—none, high

Public evaluations

Benchmarks

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

  • GPT-OSS 120B TEE
  • Mistral Small 4 119B Thinking

Intelligence Index

  • GPT-OSS 120B TEE11.6
  • Mistral Small 4 119B Thinking11.3

GPT-OSS 120B TEE tested at high reasoning effort

GPT-OSS 120B TEE: Better than 25% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • GPT-OSS 120B TEE: 30.4Top
    • Mistral Small 4 119B Thinking: 26.6

  • Math Index

    Composite of the math evaluations

    • GPT-OSS 120B TEE: 93.4
    • Mistral Small 4 119B Thinking: —

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • GPT-OSS 120B TEE: 0.2%
    • Mistral Small 4 119B Thinking: 1.0%Top

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • GPT-OSS 120B TEE: 0.0%Top
    • Mistral Small 4 119B Thinking: 0.0%Top

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • GPT-OSS 120B TEE: —
    • Mistral Small 4 119B Thinking: 241 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • GPT-OSS 120B TEE: 608 EloTop
    • Mistral Small 4 119B Thinking: 362 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • GPT-OSS 120B TEE: 4.0%Top
    • Mistral Small 4 119B Thinking: 0.0%

  • AA-LCRv1.1

    Reasoning across long documents

    • GPT-OSS 120B TEE: 52.0%Top
    • Mistral Small 4 119B Thinking: 49.7%

  • MLCR-AA

    Long-context reasoning over medical records

    • GPT-OSS 120B TEE: 1.7%Top
    • Mistral Small 4 119B Thinking: 1.7%Top

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • GPT-OSS 120B TEE: 19.6%Top
    • Mistral Small 4 119B Thinking: 9.9%

  • GPQA Diamond

    Graduate-level science questions

    • GPT-OSS 120B TEE: 78.2%Top
    • Mistral Small 4 119B Thinking: 76.9%

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • GPT-OSS 120B TEE: 80.8%
    • Mistral Small 4 119B Thinking: —

  • IFBench

    Following precise output instructions

    • GPT-OSS 120B TEE: 69.0%Top
    • Mistral Small 4 119B Thinking: 48.2%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • GPT-OSS 120B TEE: 0.0%Top
    • Mistral Small 4 119B Thinking: 0.0%Top

  • SciCode

    Scientific computing in Python

    • GPT-OSS 120B TEE: 34.0%
    • Mistral Small 4 119B Thinking: 38.8%Top

  • LiveCodeBench

    Recent competitive programming problems

    • GPT-OSS 120B TEE: 87.8%
    • Mistral Small 4 119B Thinking: —

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • GPT-OSS 120B TEE: 93.4%
    • Mistral Small 4 119B Thinking: —

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • GPT-OSS 120B TEE: 23.5%Top
    • Mistral Small 4 119B Thinking: 17.4%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • GPT-OSS 120B TEE: 65.8%Top
    • Mistral Small 4 119B Thinking: 41.2%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • GPT-OSS 120B TEE: 52.0%Top
    • Mistral Small 4 119B Thinking: 49.7%

Scores from Artificial Analysis, updated Oct 26, 2026. GPT-OSS 120B TEE was tested at high reasoning effort; compare scores measured at the same effort.Artificial Analysis

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": "TEE/gpt-oss-120b",
    "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="TEE/gpt-oss-120b",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

OpenAI

Other OpenAI models, starting with the same family.