GPT 5 Mini

A lightweight version of GPT-5 for cost-sensitive applications. Balances advanced capabilities with efficient resource usage.

Reasoning
Vision

Added Aug 7, 2025

Pricing

Auto routing · per 1M tokens

Input: $0.25
Output: $2.00
Cache read: $0.025

Compare provider prices

Specifications

Context window: 400K
Max output: 128K
Avg output (7d): 1K tokens; Longer than 71% of models

Benchmarks

Artificial Analysis

LMArena

Vectara

Design Arena

Sourced from Artificial Analysis.

Intelligence Index

16.8

Better than 59% of models compared

Coding Index

15.6

Better than 17% of models compared

Agentic Index

6.8

Better than 39% of models compared

Agentic work

AutomationBench-AA

Workflow automation with guardrail penalties

6.5%

Better than 31% of models compared

AA-Briefcase

Agentic knowledge work (Elo)

424 Elo

Better than 21% of models compared

GDPval-AA v2

Economically valuable tasks (Elo)

770 Elo

Better than 36% of models compared

Document reasoning

GDP.pdf

Professional PDF reasoning: all-pass rate

8.4%

Better than 35% of models compared

AA-LCR v1.1

Long context reasoning with updated grading

72.3%

Better than 68% of models compared

Reasoning

HLE

Humanity's Last Exam

21.5%

Better than 69% of models compared

IFBench

Instruction-following benchmark

75.4%

Better than 93% of models compared

CritPt

Research-level physics reasoning

0.0%

Coding

Terminal-Bench v4.0

Practical coding and terminal tasks

0.0%

Better than 16% of models compared

SciCode

Python programming for scientific computing

39.0%

Better than 23% of models compared

LiveCodeBench

Contamination-free coding benchmark

83.8%

Better than 95% of models compared

Math

AIME 2025

American Invitational Mathematics Examination 2025

90.7%

Better than 91% of models compared

Knowledge

MMLU-Pro

Professional and academic subject knowledge

83.7%

Better than 86% of models compared

AA-Omniscience Accuracy

Proportion of correctly answered questions

25.0%

AA-Omniscience Hallucination Rate

Rate of incorrect answers among non-correct responses

56.4%

Legacy benchmarks

GPQA Diamond (legacy)

Graduate-level scientific reasoning

82.8%

Better than 72% of models compared

Terminal-Bench Hard (legacy)

Agentic coding and terminal use

33.3%

Better than 77% of models compared

T²-Bench Telecom (legacy)

Conversational AI agents in dual-control scenarios

68.4%

Better than 61% of models compared

AA-LCR (unversioned / legacy)

Long context reasoning evaluation

72.3%

Better than 68% of models compared

GDPval-AA (unversioned / legacy)

Economically valuable tasks

12.7%

Last updated Oct 3, 2026