GPT-4o
OpenAI🇺🇸 United States
Context window128K tokens
Input / 1M tokens$2.5
Output / 1M tokens$10
Version History
gpt-4o-2025-12-17minor
GPT-4o December snapshot with improved real-time audio mode quality and enhanced JSON schema structured output compliance.
gpt-4o-2025-10-14minor
GPT-4o October snapshot with vision improvements and better performance on document understanding tasks.
gpt-4o-2025-08-06minor
GPT-4o August 2025 snapshot with improved structured output reliability and expanded multilingual performance.
gpt-4o-2024-11-20minor
November snapshot with improved instruction following and reduced verbosity.
gpt-4o-2024-08-06minor
GPT-4o August snapshot adds structured outputs (JSON Schema) and improved function calling accuracy.
gpt-4omajor
OpenAI released GPT-4o, a multimodal model delivering 2x faster inference than GPT-4 Turbo at 50% lower cost with 128K context window and improved non-English language capabilities.
Benchmark Scores
Full leaderboard →92.8%
DocVQA
53.6%
GPQA
90.2%
HumanEval
53.0%
LiveCodeBench
76.6%
MATH
88.7%
MMLU
64.0%
MMLU-Pro
75.8%
MMMU
98.0 tokens_per_sec
Speed (tok/s)
38.8%
SWE-bench Verified