GPT-4 Turbo: Complete Benchmark Performance Guide

openai turbo 2023-11-06 Proprietary

In-depth analysis of GPT-4 Turbo performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendoropenai
Versionturbo
Release Date2023-11-06
Context Window128000 tokens
Input Modalitiestext, image
Output Modalitiestext
LicenseProprietary
Documentationhttps://platform.openai.com/docs/models

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 84.4 % 2023-11-06 5-shot view
HUMANEVAL 80.5 pass@1 2023-11-06 — view
GSM8K 85.2 % 2023-11-06 0-shot CoT view
MATH 50.9 % 2023-11-06 0-shot CoT view
BBH 78.5 % 2023-11-06 3-shot CoT view
GPQA 49.2 % 2023-11-06 0-shot view
IFEVAL 77.9 % 2023-11-06 prompt_strict view
ARC 94.2 % 2023-11-06 challenge view
MUSR 61.1 % 2023-11-06 0-shot view
WINOGRANDE 80.8 % 2023-11-06 0-shot view

Pricing

Tier Price Currency
Input$10 / MtokUSD
Output$30 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://openai.com/api/pricing/ · as of 2023-11-06

Compliance

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

GPT-4 Turbo

Model Overview

OpenAI GPT-4 Turbo 模型, 128K 上下文窗口, 支持视觉输入, 相比 GPT-4 价格降低 3 倍, 性能提升。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Openai turbo 2023-11-06 128K text, image text Proprietary

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 84.4 % 5-shot
HumanEval 80.5 pass@1 —
GSM8K (Grade School Math 8K) 85.2 % 0-shot CoT
MATH 50.9 % 0-shot CoT
BBH (BIG-Bench Hard) 78.5 % 3-shot CoT
GPQA 49.2 % 0-shot
IFEval 77.9 % prompt_strict
ARC 94.2 % challenge
MUSR 61.1 % 0-shot
WinoGrande 80.8 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • MMLU score 84.4, strong knowledge reasoning.
  • HumanEval 80.5, excellent code generation.
  • GSM8K 85.2, robust math reasoning.
  • Input Modalities: text, image, audio.
  • Context window 128K.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

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

References