GPT-4 0125 Preview: Complete Benchmark Performance Guide

openai 4-0125-preview 2024-01-25 Proprietary

In-depth analysis of GPT-4 0125 Preview performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendoropenai
Version4-0125-preview
Release Date2024-01-25
Context Window128000 tokens
Input Modalitiestext
Output Modalitiestext
LicenseProprietary
Documentationhttps://platform.openai.com/docs/models

Benchmark Performance

BenchmarkScoreUnitEvaluated AtNotesSource
MMLU82.6%2024-01-255-shotview
HUMANEVAL80.3pass@12024-01-25—view
GSM8K82.9%2024-01-250-shot CoTview
MATH42.3%2024-01-250-shot CoTview
BBH78.2%2024-01-253-shot CoTview
GPQA46.8%2024-01-250-shotview
IFEVAL71.4%2024-01-25prompt_strictview
ARC95.5%2024-01-25challengeview
MUSR65.8%2024-01-250-shotview
WINOGRANDE86.6%2024-01-250-shotview

Pricing

TierPriceCurrency
Input$10 / MtokUSD
Output$30 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://openai.com/api/pricing/ · as of 2024-01-25

Compliance

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

GPT-4 0125 Preview

Modellöversikt

OpenAI GPT-4 0125 预览版, 128K 上下文, 修复函数调用重复调用问题, 改进代码生成能力。

Kärnspecifikationer

LeverantörVersionReleasedatumKontextfönsterInmatningsmodaliteterUtmatningsmodaliteterLicens
Openai4-0125-preview2024-01-25128KtexttextProprietary

Benchmarkprestanda

BenchmarkPoängEnhetAnteckningar
MMLU (Massive Multitask Language Understanding)82.6%5-shot
HumanEval80.3pass@1—
GSM8K (Grade School Math 8K)82.9%0-shot CoT
MATH42.3%0-shot CoT
BBH (BIG-Bench Hard)78.2%3-shot CoT
GPQA46.8%0-shot
IFEval71.4%prompt_strict
ARC95.5%challenge
MUSR65.8%0-shot
WinoGrande86.6%0-shot

Priser

InmatningUtmatningCacheläsningCacheskrivning
————

per miljon tokens

Styrkor

  • MMLU score 82.6, strong knowledge reasoning.
  • HumanEval 80.3, excellent code generation.
  • 上下文窗口 128K，支持长文本。

Svagheter

  • 闭源专有模型，不支持自托管。

Användningsområden

  • 代码生成与调试
  • 长文档摘要
  • Agent 工作流与工具调用

Referenser