GPT-5.4 vs GPT-5.3-Codex
GPT-5.3-Codex costs less on both: $1.05 in and $8.40 out per million tokens, against $1.50 in and $9.00 out for GPT-5.4. Both on one key at UplinkSwitch, each next to its developer's list price.
Prices · per million tokens, USD
| Tokens | GPT-5.4gpt-5.4 | GPT-5.3-Codexgpt-5.3-codex |
|---|---|---|
| Input | $1.50OpenAI lists $2.50 · 40% under | $1.05OpenAI lists $1.75 · 40% under |
| Cached input | $0.15OpenAI lists $0.25 · 40% under | $0.105OpenAI lists $0.175 · 40% under |
| Output | $9.00OpenAI lists $15.00 · 40% under | $8.40OpenAI lists $14.00 · 40% under |
| Context window | 1,050,000 tokens | 400,000 tokens |
| Max output | 128,000 tokens | 128,000 tokens |
| Input | text, image | text, image |
| Released | 2026-03-05 | 2026-02-24 |
| Formats | OpenAI SDK | OpenAI SDK |
| Developer | OpenAI | OpenAI |
List price: each developer's published rate, read October 2026. A model with no cached rate bills cached tokens at its input rate. Live from GET https://api.uplinkswitch.com/v1/catalog.
What the same work costs · USD, here and at list
| Work | GPT-5.4 | GPT-5.3-Codex |
|---|---|---|
| 1,000 chat requests | $7.50list $12.50 | $6.30list $10.50 |
| 100 long documents | $9.30list $15.50 | $6.93list $11.55 |
| A month of both, daily | $504.00list $840.00 | $396.90list $661.50 |
A chat request is 2,000 input and 500 output tokens; a long document is 50,000 in and 2,000 out; the month is 1,000 chats and 100 documents a day for 30 days, uncached. Over that month GPT-5.3-Codex costs $107.10 less here.
Same key, same URL
One base URL and one key serve both models. The request below calls GPT-5.4; put gpt-5.3-codex in the model field to call GPT-5.3-Codex instead.
shell
curl https://api.uplinkswitch.com/v1/chat/completions \
-H "authorization: Bearer sk-uplink-live-..." \
-H "content-type: application/json" \
-d '{"model":"gpt-5.4","messages":[{"role":"user","content":"Hello"}]}'Model pages: GPT-5.4, GPT-5.3-Codex. Pricing by family: GPT. More comparisons: GPT-5.4 vs Claude Opus 4.6, GPT-5.5 vs GPT-5.3-Codex.
Questions
- Which costs less, GPT-5.4 or GPT-5.3-Codex?
- GPT-5.3-Codex costs less on both: $1.05 in and $8.40 out per million tokens, against $1.50 in and $9.00 out for GPT-5.4.
- Can I switch between them by changing the model id?
- Yes. Both answer in the OpenAI SDK format at the same base URL with the same key. Change the model id from gpt-5.4 to gpt-5.3-codex and nothing else.
- Is there a subscription?
- No. Prepaid credit from $10, with no subscription or monthly fee. Each top-up locks your prices for 30 days.
- Is GPT-5.4 better than GPT-5.3-Codex?
- Better depends on the job; this page compares what the catalog can measure. GPT-5.3-Codex costs less on both: $1.05 in and $8.40 out per million tokens, against $1.50 in and $9.00 out for GPT-5.4. GPT-5.4 takes 1,050,000 tokens of context against 400,000 for GPT-5.3-Codex. GPT-5.4 is the newer release. Both run on the same key, so the fair test is your own prompt sent to each model id.