DataLearner logo
GP

GPT OSS 120B

Reasoning modelGPT OSSGPT OSS 120B

GPT Opensources 120B

Release date: 2025-08-06Updated: 2026-07-17Views: 2,169
Live demoGitHubHugging FaceCompare
Parameters
11.7B
Context length
128K
Multilingual
Supported
Reasoning ability
3/5

GPT OSS 120B is a reasoning model from OpenAI, released on 2025-08-06. It accepts text input and returns text output. The cataloged parameter count is 11.7B, with 5.1B active parameters per inference. The recorded context window is 128K, and the recorded maximum output is 128K. Cataloged capabilities include Reasoning model and Multilingual. The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

GPT OSS 120B

Model basics

Reasoning traces
Supported
Thinking modes
Standard Mode (Default)Thinking Mode
Context length
128K tokens
Max output length
128K tokens
Model type
Reasoning model
Modality (in / out)
Text → Text
Release date
2025-08-06
Model file size
240GB
MoE architecture
Yes
Total params / Active params
11.7B / 5.1B
Knowledge cutoff
No data
GPT OSS 120B

Open source & experience

Code license
Apache 2.0
Weights license
Apache 2.0- Commercial use permitted
GitHub repo
github.com/openai/gpt-oss
Hugging Face
huggingface.co/openai/gpt-oss-120b
Live demo
N/A
GPT OSS 120B

Official resources

Paper
Paper
DataLearnerAI blog
DataLearnerAI blog
GPT OSS 120B

API details

API speed
2/5
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Learn about pricing modes
Standard
TypeConditionInputOutput
Text-$0.350/ 1M tokens$0.750/ 1M tokens

“—” means the modality is not billed in that direction, or the vendor has not published a price for it.

GPT OSS 120B

Benchmark Results

GPT OSS 120B currently shows benchmark results led by AIME 2024 (2 / 62, score 96.60), MMLU (11 / 124, score 90), AIME2025 (17 / 107, score 97.90). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage

General Knowledge

6 evaluations
Benchmark / mode
Score
Rank/total
MMLU
Thinking Mode
90
11 / 124
MMLU Pro
Thinking Mode
79
67 / 134
LiveBench
Standard Mode
46.09
104 / 117
AA Intelligence Index (historical versions)
High
24.13
27 / 28
HLE
Thinking Mode
14.90
158 / 197
HLE
Thinking ModeTools
19
145 / 197

General Evaluation

1 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Thinking Mode
80.10
146 / 274

Coding and Software Engineer

4 evaluations
Benchmark / mode
Score
Rank/total
CodeForces
Thinking Mode
2463
11 / 21
CodeForces
Thinking ModeTools
2622
9 / 21
SWE-bench Verified
Thinking Mode
60.10
83 / 116
SciCode
High
38.89
14 / 16

Math and Reasoning

3 evaluations
Benchmark / mode
Score
Rank/total
AIME2025
Thinking Mode
83
52 / 107
AIME2025
Thinking ModeTools
97.90
17 / 107
AIME 2024
Thinking ModeTools
96.60
2 / 62

Writing and Creative Capabilities

1 evaluations
Benchmark / mode
Score
Rank/total
Creative Writing
Standard Mode
959.20
89 / 106

Common Sense Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
SimpleBench
Thinking Mode
22.10
87 / 94

Agent Level Benchmark

2 evaluations
Benchmark / mode
Score
Rank/total
Aider-Polyglot
High
41.80
38 / 59
τ³-Banking
HighTools
12.78
13 / 13

Instruction Following

1 evaluations
Benchmark / mode
Score
Rank/total
IF Bench
Standard Mode
69
21 / 36

Long Context

1 evaluations
Benchmark / mode
Score
Rank/total
AA-LCR
High
51
29 / 29

Claw-style Agent Evaluation

1 evaluations
Benchmark / mode
Score
Rank/total
Pinch Bench
Thinking ModeTools
60.60
36 / 38

Productivity Knowledge

1 evaluations
Benchmark / mode
Score
Rank/total
GDPval-AA v2
HighTools
802.63
26 / 27
View benchmark analysisCompare with other models
GPT OSS 120B

Publisher

GPT Opensources 120B

Model Overview

GPT OSS 120B is a reasoning model from OpenAI, released on 2025-08-06.

It accepts text input and produces text output. Its cataloged capabilities include Reasoning model and Multilingual. The cataloged parameter count is 11.7B, with 5.1B active parameters per inference. The recorded context window is 128K, and the recorded maximum output is 128K.

The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. The page records 2 API pricing rules from Cerebras Systems; current provider pricing and conditions should be checked before deployment. The evaluation section contains 16 cataloged benchmark results with their recorded modes and scores. The page links 4 release, model-card, repository, or provider references for checking the underlying claims. Specifications, availability, and prices can change; undisclosed values are intentionally left unstated.

GPT OSS 120B

FAQ

What is GPT OSS 120B?

GPT OSS 120B is a reasoning model from OpenAI, released on 2025-08-06. It accepts text input and returns text output. The cataloged parameter count is 11.7B, with 5.1B active parameters per inference. The recorded context window is 128K, and the recorded maximum output is 128K. Cataloged capabilities include Reasoning model and Multilingual. The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

What input and output modalities does GPT OSS 120B support?

The current model record lists text as input and text as output.

What are the main recorded specifications for GPT OSS 120B?

The cataloged parameter count is 11.7B, with 5.1B active parameters per inference. The recorded context window is 128K, and the recorded maximum output is 128K. Fields without a source-backed value remain undisclosed.

Does GPT OSS 120B have API pricing?

The page records 2 API pricing rules from Cerebras Systems; current provider pricing and conditions should be checked before deployment.

Are benchmark results available for GPT OSS 120B?

The evaluation section contains 16 cataloged benchmark results with their recorded modes and scores. Compare only results that use the same benchmark version and evaluation mode.

Is GPT OSS 120B open source?

The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Review the linked license text before commercial or derivative use.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code