OpenAI · Open weights
GPT OSS 120B
An open-weight, 117B-parameter Mixture-of-Experts (MoE) language model designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized to run on a single H100 GPU with native MXFP4 quantization. The model supports configurable reasoning depth, full chain-of-thought access, and native tool use, including function calling, browsing, and structured output generation.
- Reasoning
- Tools
- Structured output
- Input
- Text
- Output
- Text
- Discovery tier
- Open weights
Pricing
USD per 1M tokens- Input
- $0.03
- Output
- $0.15
- Cache read
- $0.03
Reference prices as published by the model provider. API example
Specifications
- Context window
- 128K
- Max output
- 16.4K
- Average output (7 days)
- 1.1K tokens
- Longer than 74% of 468 models
Side by side
GPT OSS 120B compared
Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.
| GPT OSS 120BOpenAIThis model | GPT OSS 20BOpenAIRemove | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Discovery tier | Open weights | Open weights |
| Added | Feb 3, 2026 | Aug 5, 2025 |
| Limits | ||
| Context window | 128K | 131.07KBest |
| Max output | 16.38K | 16.38K |
| Average output | 1.1K tokens | 1.9K tokens |
| Pricing, USD per 1M tokens | ||
| Input | $0.03 | $0.02Best |
| Output | $0.15 | $0.10Best |
| Cache read | $0.03 | — |
| Performance | ||
| Throughput | 625.0 tok/s | — |
| Time to first token | 0.8 s | — |
| Request success (24h) | 99.8%Best | 99.6% |
| Benchmark indexes | ||
| Intelligence Index | 11.6Best | 9.0 |
| Coding Index | 30.4Best | 20.7 |
| Math Index | 93.4Best | 89.3 |
| Capabilities | ||
| Reasoning | Yes | Yes |
| Tools | Yes | Yes |
| Structured output | Yes | No |
| Reasoning effort | — | — |
Public evaluations
Benchmarks
Independent evaluation results, with how each score ranks among the models in the Noumen catalog.
- GPT OSS 120B
- GPT OSS 20B
Intelligence Index
- GPT OSS 120B11.6
- GPT OSS 20B9.0
GPT OSS 120B tested at high reasoning effort
GPT OSS 120B: Better than 25% of 326 scored models
Indexes
Coding Index
Composite of the coding evaluations
- GPT OSS 120B: 30.4Top
- GPT OSS 20B: 20.7
Math Index
Composite of the math evaluations
- GPT OSS 120B: 93.4Top
- GPT OSS 20B: 89.3
Agentic work
AutomationBench-AA
Workflow automation, penalized for guardrail violations
- GPT OSS 120B: 0.2%Top
- GPT OSS 20B: 0.2%
AutomationBench-AA tasks completed
Workflows finished end to end without a guardrail violation
- GPT OSS 120B: 0.0%Top
- GPT OSS 20B: 0.0%Top
GDPval-AAv2.1
Economically valuable occupational tasks, rated head to head
- GPT OSS 120B: 608 EloTop
- GPT OSS 20B: 339 Elo
Document reasoning
GDP.pdf
Professional reasoning over PDF documents, all-pass rate
- GPT OSS 120B: 4.0%Top
- GPT OSS 20B: 2.0%
AA-LCRv1.1
Reasoning across long documents
- GPT OSS 120B: 52.0%Top
- GPT OSS 20B: 34.7%
MLCR-AA
Long-context reasoning over medical records
- GPT OSS 120B: 1.7%Top
- GPT OSS 20B: 1.1%
Reasoning and knowledge
Humanity’s Last Exam
Expert-level questions across academic fields
- GPT OSS 120B: 19.6%Top
- GPT OSS 20B: 11.0%
GPQA Diamond
Graduate-level science questions
- GPT OSS 120B: 78.2%Top
- GPT OSS 20B: 68.8%
MMLU-Pro
Multi-subject knowledge and reasoning
- GPT OSS 120B: 80.8%Top
- GPT OSS 20B: 74.8%
IFBench
Following precise output instructions
- GPT OSS 120B: 69.0%Top
- GPT OSS 20B: 65.1%
Coding
Terminal-Benchv4.0
Practical software engineering in a terminal
- GPT OSS 120B: 0.0%Top
- GPT OSS 20B: 0.0%Top
SciCode
Scientific computing in Python
- GPT OSS 120B: 34.0%
- GPT OSS 20B: 38.9%Top
LiveCodeBench
Recent competitive programming problems
- GPT OSS 120B: 87.8%Top
- GPT OSS 20B: 77.7%
Math
AIME 2025
Competition mathematics, 2025 problems
- GPT OSS 120B: 93.4%Top
- GPT OSS 20B: 89.3%
Legacy evaluations
Terminal-Bench Hard
Earlier terminal task suite
- GPT OSS 120B: 23.5%Top
- GPT OSS 20B: 10.6%
τ²-Bench Telecom
Tool use in a customer-support simulation
- GPT OSS 120B: 65.8%Top
- GPT OSS 20B: 60.2%
AA-LCR (unversioned)
Long-context reasoning, earlier grading
- GPT OSS 120B: 52.0%Top
- GPT OSS 20B: 34.7%
Scores from Artificial Analysis, updated Oct 26, 2026. GPT OSS 120B 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.
- Throughput
- 625.0 tok/s
- Faster than 98% of 263 models
- Time to first token
- 0.8 s
- Quicker than 98% of 263 models
- Prompt cache hits
- 7.1%
- About $0.03 per 1M input tokens with caching
- Request success (24h)
- 99.8%
- 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.
- 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.
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-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="openai/gpt-oss-120b",
messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)OpenAI
Related models
Other OpenAI models, starting with the same family.