GPT-4o vs GPT-5: pricing & cost comparison
On input tokens, GPT-5 is the cheaper of the two — 50% less per million ($2.5 vs $1.25). Both charge $10 per million output tokens.
Side by side
| GPT-4o | GPT-5 | |
|---|---|---|
| Input / 1M tokens | $2.5 | $1.25 |
| Output / 1M tokens | $10 | $10 |
| Context window | 128,000 | 400,000 |
| Token-count accuracy | exact | exact |
| Cost — 10,000 input + 2,000 output tokens | $0.045 | $0.0325 |
What a real request costs
Take a representative turn — 10,000 input + 2,000 output tokens. GPT-4o comes to $0.045, GPT-5 to $0.0325. Across 100,000 requests that's a $1250 swing in favour of GPT-5. To run the numbers on your actual prompt, paste it into the calculator and toggle Compare across all models.
Same vendor — an easy switch
Moving between GPT-4o and GPT-5 is a same-vendor move — one SDK, one tokenizer, so no recalibration. That makes this a straight cost/quality call: send the routine bulk of traffic to GPT-5 ($1.25/$10 per 1M) and reserve GPT-4o for the requests that visibly need its extra headroom. On the worked example above that split is worth about 28% per request — small per call, real money once you multiply by volume.
See the full breakdown on the dedicated pages for GPT-4o and GPT-5.
FAQ
- Is GPT-4o or GPT-5 cheaper?
- For a typical request (10,000 input + 2,000 output tokens), GPT-5 is cheaper — about 28% less, or roughly $1250 saved per 100,000 requests. GPT-4o runs $2.5/$10 per 1M input/output tokens; GPT-5 runs $1.25/$10.
- Which has the larger context window?
- GPT-5, at 400,000 tokens versus 128,000.
- How accurate are these token counts?
- GPT-4o: Exact tokenization via the canonical OpenAI vocab (o200k_base). GPT-5: Exact tokenization via the canonical OpenAI vocab (o200k_base). The dollar math itself is exact once the token count is known.
Both prices are computed from tokenmath's verified pricing table. Rates sourced from openai.com, verified 2026-05-18. Vendor pricing changes often — confirm before you commit.