NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/GPT-5
OpenAI logo

GPT-5

by OpenAI · Released Aug 2025

Multimodal
60.4
avg score
Rank #84
Compare
Better than 73% of all models
Context
400K tokens (~200 books)
Input $/1M
$1.25
Output $/1M
$10.00
Type
multimodal
License
Proprietary
Benchmarks
41 tested
Data as of
About

GPT-5 is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience. It is optimized for complex tasks that require step-by-step reasoning, instruction following, and accuracy...

Tested on 41 benchmarks · BenchGecko score 60.4. Top scores: MATH level 5 (98.1%), Fiction.LiveBench (97.2%), OTIS Mock AIME 2024-2025 (91.4%).

Looking for similar performance at lower cost?
DeepSeek V4 Flash 0731 scores 60.1 (100% as good) at $0.02/1M input · 99% cheaper
Capabilities
coding
57.3
#60 globally
reasoning
52.3
#75 globally
math
53.8
#106 globally
knowledge
61.3
#49 globally
agentic
18.3
#49 globally
language
87.5
#27 globally
general
35.4
#99 globally
Benchmark Scores
Compare All
Tested on 41 benchmarks · Ranked across 7 categories
Score Distribution (all 312 models)
0255075100
▲ You are here
codingCompare coding →
Aider polyglot

Multi-language code editing from Aider. Tests editing ability across Python, JavaScript, TypeScript, Java, C++, Go, Rust, and more.

88.0·
SWE-Bench verified

Real-world software engineering tasks from GitHub issues. Models must diagnose bugs and write patches that pass test suites. Human-verified subset of SWE-bench.

73.5·
SWE-Bench Verified (Bash Only)

SWE-bench Verified solved using only bash commands, no specialized frameworks. Tests raw terminal-based problem solving.

65.0·
reasoningCompare reasoning →
HELM — WildBench

Stanford HELM WildBench evaluation. Tests reasoning on challenging real-world tasks.

85.7·
ARC-AGI

Abstraction and Reasoning Corpus. Tests fluid intelligence through novel visual pattern recognition puzzles. Core measure of general intelligence.

65.7·
SimpleBench

Deceptively simple questions that humans find easy but AI models often get wrong. Tests common sense and reasoning gaps.

48.0·
mathCompare math →
MATH level 5

Competition-level math from AMC, AIME, and olympiad problems. Level 5 is the hardest tier, requiring creative problem-solving.

98.1·
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

91.4·
HELM — Omni-MATH

Stanford HELM evaluation of mathematical reasoning across diverse problem types.

64.7·
Quick compare:vs internlm-20bGoogle DeepMind logovs Gemini 3.6 FlashDeepSeek logovs DeepSeek V4 Flash 0731
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Model Family · OpenAI GPT-5
Aug 2025
60.4
$1.25/M in400Kctx41 benchmarks
Aug 2025
$0.63/M in(-0.63)400Kctx
Aug 2025
89.0+28.6
$1.25/M in(+0.63)128Kctx(-272K)7 benchmarks
Sep 2025
$1.25/M in400Kctx(+272K)
Sep 2025
$0.63/M in(-0.63)400Kctx
Oct 2025
$10.00/M in(+9.38)400Kctx
Oct 2025
$2.50/M in(-7.50)400Kctx
Aug 2025
56.0-33.0
$0.25/M in(-2.25)400Kctx36 benchmarks
Aug 2025
$0.13/M in(-0.13)400Kctx
Aug 2025
37.7-18.3
$0.05/M in(-0.07)400Kctx33 benchmarks
Aug 2025
$0.03/M in(-0.03)400Kctx
Oct 2025
48.7+11.0
$15.00/M in(+14.97)400Kctx10 benchmarks
Oct 2025
$7.50/M in(-7.50)400Kctx
See the full GPT-5 family →
Similar Models
internlm-20b
Unknown
60.3n/a
Google DeepMind logo
Gemini 3.6 Flash
Google DeepMind
60.6$0.75/1M
DeepSeek logo
DeepSeek V4 Flash 0731
DeepSeek
60.1$0.02/1M
Links
Info
OpenAIPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@OpenAI
BenchGecko API
gpt-5
Specifications
  • Typemultimodal
  • Context400K tokens (~200 books)
  • ReleasedAug 2025
  • LicenseProprietary
  • StatusActive
  • Cost / Message~$0.013
Available On
OpenAI logoOpenAI$1.25
Categories
codingreasoningmathknowledgeagenticlanguagegeneral
Powers Agents
OpenHands71.8%View all agents →
Learn More
context-windowtransformertokens
Share & Export
Tweet
Related Models
internlm-20b
60.3
Google DeepMind logoGemini 3.6 Flash
60.6
DeepSeek logoDeepSeek V4 Flash 0731
60.1
Stable Beluga 2
60.1
GPT-5 is a proprietary multimodal AI model by OpenAI, released in August 2025. It has an average benchmark score of 60.4. Context window: 400K tokens.

Related Models

internlm-20b · UnknownGemini 3.6 Flash · Google DeepMindDeepSeek V4 Flash 0731 · DeepSeekStable Beluga 2 · UnknownQwen-14B · Alibaba Qwen

Benchmarks

MATH level 5Fiction.LiveBenchOTIS Mock AIME 2024-2025Aider polyglotHELM — IFEval

Related Pages

OpenAI · ProviderOpenAI · EconomyAll ModelsCompare ModelsPricingDevelopers · APIContext Window · Glossary

Key facts · as of 2026-10-05

  • GPT-5 by OpenAI. BenchGecko score 60.4, rank 84 of 312 scored models (normalized average of public benchmark scores).
  • List price $1.25 input · $10.00 output per 1M tokens (as of 2026-10-05).
  • Sold by 3 providers (as of 2026-10-05): Azure $1.25 in / $10.00 out · OpenAI $1.25 in / $10.00 out · Azure $1.38 in / $11.00 out. Every provider

How to cite · data as of 2026-10-05

GPT-5 · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/gpt-5

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP