OpenAI ChatGPT Costs

2025-07-15

(All prices are per 1,000 tokens – abbreviated as $/KToken)

🔹 GPT-4o / GPT-4o-mini (latest, fast, and capable)

• Prompt tokens: $5.00 per 1,000,000 tokens → $0.005 / KToken
• Output tokens: $15.00 per 1,000,000 tokens → $0.015/KToken

🔹 GPT-4 Turbo (128k context)

• Prompt tokens: $10.00 per 1,000,000 tokens → $0.01 / KToken
• Output tokens: $30.00 per 1,000,000 tokens → $0.03 / KToken

🔹 GPT-3.5 Turbo

• Both prompt and output: $0.50 – $1.00 per 1,000,000 tokens → $0.0005 – $0.001 / KToken

🔢 Capability and Cost Comparison

Model	Capability (⭐ = relative power)	Prompt Cost (\(/KToken) | Output Cost (\)/KToken)
GPT-4o	⭐⭐⭐⭐ (very fast & capable)	$0.005	$0.015
GPT-4 Turbo	⭐⭐⭐⭐ (very capable)	$0.01	$0.03
GPT-3.5 Turbo	⭐⭐ (fast, good for many tasks)	$0.0005 – $0.001	Same as prompt

🧠 Summary

• GPT-4o is the best value right now — it’s cheaper and faster than GPT-4 Turbo, with similar or better performance in most tasks.
• GPT-4 Turbo is still powerful and offers a long 128k context window.
• GPT-3.5 Turbo is the cheapest and useful for lighter tasks, though much less capable than the GPT-4 family.

🔎 Notes

• “Tokens” are chunks of text—roughly 3 to 4 words per token.
• Total cost = cost of prompt tokens + cost of output tokens.
• Always check OpenAI’s Pricing Page and your API dashboard for the latest rates.