Last verified
STILL SOLDLEGACY TIER128KTEXT + VISIONPROMPT CACHING

GPT-4o API Pricing

GPT-4o is two generations old and still on OpenAI's price list: $2.50/M input, $1.25/M cached input and $10/M output, an effective $2.16/M on this site's agentic blend. Two things are worth knowing before you treat it as an archive. It is not the model OpenAI is retiring on October 23, 2026 — that is the frozen gpt-4o-2024-05-13 snapshot, which still lists at the launch price of $5/$15 while the plain gpt-4o alias has charged half that since August 2024. And on our blend it undercuts GPT-5.5, the model most migration guides point to, because GPT-5.5 bills output at $30.

Input - per 1M tokens
$2.50/M
Source OpenAI half of launch
Output - per 1M tokens
$10.00/M
Was $15 at launch cut Aug 2024
Cached input
$1.25/M
Not the 90% newer tiers get 50% off
Effective - agentic blend
$2.16/M
92/8 split - 82% cache
§ 01 / TERMINAL

Run the numbers.

Rates are the current ones for the gpt-4o alias, not the frozen 2024-05-13 snapshot. Cache misses bill at the full input rate here: OpenAI publishes no cache-write charge for this model, so there is no separate write line to model.

$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
Words equivalent (English)
Effective rate
Open full calculator (all models · share URL · CSV) →
§ 02 / SCENARIOS

Real-world presets.

§ 03 / TOKENIZER

Paste text. See tokens. See cost.

Estimate · tiktoken-cl100k_base · ≈4 chars/token Auto-counts as you type

This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.

Characters 640
Words 112
Tokens (estimated) 160 tokens
Cost as input · uncached $0.0004 USD
Cost as output · uncached $0.0016 USD
Cost as cached input $0.0002 USD
§ 04 / SHELF

Up against the shelf.

All models →
Model Input /M Output /M Effective blended Context Best for
GPT-4o Current $2.50 cache $1.25 $10.00 $2.16 current page 128K Legacy multimodal apps already wired to it
GPT-5.4 mini $0.75 cache $0.075 $4.50 $0.541 cheaper 400K The honest replacement - newer and a quarter the blend
GPT-5.4 $2.50 cache $0.25 $15.00 $1.80 cheaper 1.05M Same input rate, newer model, bigger context
GPT-5.5 $5.00 cache $0.50 $30.00 $3.61 pricier on the blend 1M Named as the migration target, but costs more here
Claude Sonnet 4.6 $3.00 cache $0.30 $15.00 $2.05 about level 1M Legacy Claude production baselines
Gemini 2.5 Pro $1.25 cache $0.125 $10.00 $1.10 cheaper 2M Gemini long-context work
DeepSeek V4 Pro $1.32 cache $0.044 $3.96 $0.569 cheaper 1M Low-cost reasoning and coding
§ 05 / DEEP LINKS

Specific scenarios.

All calculators →

Frequently asked.

What the alias actually costs, why it is not the id being retired, and what to move to if you are moving.

Q · 01 Is GPT-4o still available? +
Yes. gpt-4o is on OpenAI's current price table and in its model catalogue, described there as a fast, intelligent, flexible GPT model. It appears on no retirement table. What is being retired on October 23, 2026 is the frozen snapshot gpt-4o-2024-05-13. The distinction is easy to miss and we missed it ourselves until September 6, 2026 — but OpenAI's retirement table is explicit about aliases where they are covered, listing gpt-4-0613 | gpt-4 and gpt-4-turbo | gpt-4-turbo-2024-04-09, and the gpt-4o row names no alias at all.
Q · 02 How much does GPT-4o cost? +
$2.50/M input, $1.25/M cached input and $10/M output on the standard tier. Batch halves it to $1.25/$5. Fast mode is $4.25/$2.125/$17 — a 1.7x multiplier, not the 2x the GPT-5.x and GPT-6 rows carry, which is a reminder that the tier multiplier is per-model and has to be read rather than assumed. There is no Flex row for this model.
Q · 03 Why do I see $5 and $15 quoted elsewhere? +
Because that was the launch price, and because the snapshot that launched at it is still sold. OpenAI cut the alias to $2.50/$10 on August 6, 2024 with the gpt-4o-2024-08-06 release — its Structured Outputs announcement says developers save 50% on inputs ($2.50/1M input tokens) and 33% on outputs ($10.00/1M output tokens) compared to gpt-4o-2024-05-13. The dated snapshot kept the old rates and still lists at $5/$15 today. Any page quoting $5/$15 for plain gpt-4o is two years out of date; this one was, until September 6, 2026.
Q · 04 Is cached input a 90% discount like the newer models? +
No. GPT-4o caches at half the input rate — $1.25 against $2.50 — where the GPT-5 and GPT-6 families cache at a tenth. On a cache-heavy workload that difference matters more than the sticker: it is why GPT-4o's effective blended rate is $2.16 while GPT-5.4, which shares the identical $2.50 input rate, lands at $1.80.
Q · 05 What should teams use instead? +
GPT-5.4 mini for most of it — newer, a 400K context against 128K, and about a quarter of the blended rate. GPT-5.4 if you want the same $2.50 input rate on a much newer model with a 1.05M context. Note what the shelf below shows about the usual advice: GPT-5.5 is frequently named as the GPT-4o replacement, but at $30/M output it is 1.7x more expensive than GPT-4o on our blend, not cheaper.