Back to Models

GPT-4 (Original)

OpenAI
Released: 2023-03-14
Type: Frontier Model
Context
8k
Max Out
4.1k
Cutoff
2021-09
Params
Unknown (MoE)

Evolution: What Changed?

  • Massive reasoning jump
  • Multimodal inputs (vision)
  • Reduced hallucination
Compared to gpt-3-5-turbo

Evolution: What Changed?

  • Massive reasoning jump
  • Multimodal inputs (vision)
  • Reduced hallucination
compared to previous generation

The Breakdown

The GPT-4 (Original) represents a significant leap forward in OpenAI's lineup. Released in 2023-03-14, it targets developers and enterprise use cases with a specific focus on reasoning capabilities over raw conversational speed. While previous iterations in the Frontier Model category often struggled with complex multi-step instruction following, this model introduces a refined architecture that dramatically improves adherence to system prompts and reduces hallucination rates in technical domains. It competes directly with top-tier frontier models but carves out a distinct niche for workflows where precision and context retention matter more than creative flair. For businesses looking to integrate reliable AI agents, GPT-4 (Original) offers a compelling balance of performance and cost-efficiency.

Overall Score
9.2
/10
Pricing (per 1k tokens)
Input$0.03
Output$0.06
Currency: USD

The Good

  • First model to show true reasoning
  • Massive leap over GPT-3.5
  • Highly reliable instruction following

The Bad

  • Extremely slow
  • Very expensive

The Verdict

The OG smart model. It changed the world, but later Turbo versions made it faster and cheaper.

Performance Benchmarks

math
70
human eval
67
mmlu
86.4