gpt-4.1-mini

by openai

Pricing

Input $0.04 / 1M tokens
Output $0.16 / 1M tokens

GPT‑4.1 Mini Release by OpenAI


🗓️ Overview

GPT‑4.1 Mini launched on April 14, 2025, alongside GPT‑4.1 and GPT‑4.1 Nano.
It offers a 1 million-token context window, optimized for a balance of performance, speed, and affordability.


✨ Features

  • Multimodal: Supports both text and image input
  • Context Window: Up to 1,047,576 tokens
  • Output Limit: Up to ~32K tokens per generation
  • Low Latency: Faster response times than the full GPT‑4.1 model
  • High Instruction-Following Accuracy: Strong performance in tasks requiring understanding and structured responses

📊 Benchmarks

  • MMLU (Reasoning): ~82%
  • GPQA (Factual QA): ~51.1%
  • HumanEval (Coding): ~87.2%
  • MMMU (Multimodal Tasks): ~59.4%

Scores based on OpenAI test results and community benchmarking.
Outperforms Claude Haiku and approaches GPT‑4 Turbo performance in reasoning and multimodal tasks.


💰 Pricing

Token TypeCost per 1M Tokens
Input$0.40
Cached Input$0.10
Output$1.60

Fine-Tuning (Optional)

  • Input: $0.80
  • Cached Input: $0.20
  • Output: $3.20
  • Training: $5.00 per 1M tokens

Pricing from OpenAI API Pricing


🎯 Use Cases

  • Mid-scale coding assistance
  • Document summarization & Q&A systems
  • Image + text multimodal understanding
  • Customer support bots
  • Embedded AI assistants in mobile/web apps

🚪 Availability

  • API + Playground: Since April 14, 2025
  • ChatGPT: Rolled out as a fallback/default for free-tier users starting May 14, 2025
  • ChatGPT Plus/Pro: Users may be upgraded to GPT‑4.1 or GPT‑4.1 Turbo based on load

⚠️ Limitations

  • Slightly weaker than full GPT‑4.1 on deep reasoning and very large prompts
  • No native audio or video support
  • Available via API and in ChatGPT (as fallback), not user-selectable

🔍 Comparison Table

FeatureGPT‑4.1GPT‑4.1 MiniGPT‑4.1 NanoGPT‑4o Mini
Context Window1M tokens1M tokens1M tokens128K tokens
Input Price (1M)$2.00$0.40$0.10~$0.15
Output Price (1M)$8.00$1.60$0.40~$0.60
LatencyMediumLowVery LowMedium
Vision QualityHighHighMidHigh

🧠 Summary

GPT‑4.1 Mini hits the sweet spot between power and cost.
It’s great for real-world AI agents, multimodal tools, and coding tasks where GPT‑4.1 would be too expensive and Nano not quite enough.

Highly recommended as the default LLM for most mid-range applications.