GPT-4o (2024-05-13): Complete Benchmark Performance Guide

openai 4o-2024-05-13 2024-05-13 Proprietary

In-depth analysis of GPT-4o (2024-05-13) performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendoropenai
Version4o-2024-05-13
Release Date2024-05-13
Context Window128000 tokens
Input Modalitiestext, image, audio
Output Modalitiestext, audio
LicenseProprietary
Documentationhttps://platform.openai.com/docs/models

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 89.8 % 2024-05-13 5-shot view
HUMANEVAL 83.2 pass@1 2024-05-13 — view
GSM8K 92.6 % 2024-05-13 0-shot CoT view
MATH 67.1 % 2024-05-13 0-shot CoT view
BBH 86.7 % 2024-05-13 3-shot CoT view
GPQA 58.8 % 2024-05-13 0-shot view
IFEVAL 86.7 % 2024-05-13 prompt_strict view
ARC 96.8 % 2024-05-13 challenge view
MUSR 62.8 % 2024-05-13 0-shot view
WINOGRANDE 89.3 % 2024-05-13 0-shot view

Pricing

Tier Price Currency
Input$5 / MtokUSD
Output$15 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://openai.com/api/pricing/ · as of 2024-05-13

Compliance

  • Data Residency: US
  • SOC2: ✓
  • HIPAA: ✓
  • GDPR: ✓
  • ISO 27001: ✓

GPT-4o (2024-05-13)

Model Overview

OpenAI GPT-4o 2024-05-13 快照版本, 多模态旗舰模型, 支持文本/图像/音频输入输出。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Openai 4o-2024-05-13 2024-05-13 128K text, image, audio text, audio Proprietary

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 89.8 % 5-shot
HumanEval 83.2 pass@1 —
GSM8K (Grade School Math 8K) 92.6 % 0-shot CoT
MATH 67.1 % 0-shot CoT
BBH (BIG-Bench Hard) 86.7 % 3-shot CoT
GPQA 58.8 % 0-shot
IFEval 86.7 % prompt_strict
ARC 96.8 % challenge
MUSR 62.8 % 0-shot
WinoGrande 89.3 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • MMLU score 89.8, strong knowledge reasoning.
  • HumanEval 83.2, excellent code generation.
  • GSM8K 92.6, robust math reasoning.
  • Input Modalities: text, image, audio.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Code generation and debugging
  • Vision and image understanding
  • Agent workflows and tool use

References