GPT-5 High Benchmark

OpenAI's most advanced model analyzed across comprehensive AI benchmarks

Released: 8/7/2025by OpenAIProprietary Model
66.7
Intelligence Index
#1 globally
54.9
Coding Index
#15 globally
97.5
Math Index
#1 globally
125.5
Output Speed

Key Performance Metrics

87.1%
MMLU-Pro Score
Knowledge & Reasoning
85.4%
GPQA Diamond
Scientific Reasoning
66.8%
LiveCodeBench
Coding Performance
125.5
Output Speed
Tokens per Second
$3.44
Per 1M tokens
Pricing

Token Speed Simulator

Compare real-time token generation speeds. Simulating 200 tokens (≈150 words) to show how different speeds affect text generation.

GPT-5 (high)

125.5 tokens/s
Progress0/200 tokens

Gemini 2.5 Flash-Lite (Reasoning)

526.0 tokens/s
Progress0/200 tokens
Simulating 200 tokens (≈150 words) generation • Based on median output tokens per second from analysis data

GPT-5 High Performance Analysis

Key Strengths

  • Exceptional Mathematical Reasoning: 97.5 Math Index score places it among the top mathematical reasoning models
  • Superior General Intelligence: 66.7 Intelligence Index demonstrates world-class reasoning capabilities
  • Advanced Coding Performance: 54.9 Coding Index shows excellent software development abilities
  • Scientific Reasoning Excellence: 85.4% GPQA Diamond score in graduate-level science questions

Benchmark Details

Global Intelligence Rank:#1
Coding Performance Rank:#15
Math Reasoning Rank:#1
Cost Efficiency:$3.44/1M tokens

Experience GPT-5 High's Capabilities

Join thousands of users who are already leveraging GPT-5 High's world-class performance in reasoning, coding, and mathematical problem-solving through our unified AI platform.