OpenAI logo

GPT-5 mini

Multimodal
OpenAI

A faster and more cost-effective version of GPT-5 for well-defined tasks. Excellent for clearly formulated tasks and precise prompts with strong reasoning capabilities at reduced cost.

Key Specifications

Parameters
-
Context
400.0K
Release Date
August 7, 2025
Average Score
60.0%

Timeline

Key dates in the model's history
Announcement
August 7, 2025
Today
March 26, 2026

Technical Specifications

Parameters
-
Training Tokens
-
Knowledge Cutoff
May 30, 2024
Family
-
Capabilities
MultimodalZeroEval

Pricing & Availability

Input (per 1M tokens)
$0.25
Output (per 1M tokens)
$2.00
Max Input Tokens
400.0K
Max Output Tokens
128.0K
Supported Features
Function CallingStructured OutputCode ExecutionWeb SearchBatch InferenceFine-tuning

Benchmark Results

Model performance metrics across various tests and benchmarks

Reasoning

Logical reasoning and analysis
GPQA
GPT-5 mini - Diamond thinking without tools This method analysis demonstrates work GPT-5 mini, mode thinking "Diamond" without application tools. Approach on abilities model to reasoning and In this method model: - analysis tasks with verification - chain reasoning - problem with different for process - evidence Results show, that even without access to "Diamond" thinking significantly improves performance GPT-5 mini in complex tasks, requiring Especially improvements in tasks evidence, analysis and logical By comparison with this method accuracy on 27% and analysis on 34% in tests FrontierMath and GPQASelf-reported
82.3%

Other Tests

Specialized benchmarks
AIME 2025
GPT-5 mini with included mode thinking (without tools) - mathematicsSelf-reported
91.1%
FrontierMath
GPT-5 mini with included mode thinking (only with tool python) - FrontierMath levels 1-3 mathematicsSelf-reported
22.1%
Humanity's Last Exam
GPT-5 mini with mode thinking (without tools) - set questions expert level by various subjectsSelf-reported
16.7%
HMMT 2025
GPT-5 mini with included mode thinking (without tools) - Harvard-MIT Mathematics Tournament.Self-reported
87.8%

License & Metadata

License
proprietary
Announcement Date
August 7, 2025
Last Updated
July 24, 2025

Similar Models

All Models

Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.