GPT-5.4 pricing & benchmarks
GPT-5.4 is available from OpenAI at $2.50 per million input tokens and $15.00 per million output tokens ($5.63 blended at 3:1). It accepts up to 1,050,000 tokens of context. GPT-5.4 is OpenAI’s latest frontier model, unifying the Codex and GPT lines into a single system. It features a 1M+ token context window (922K input, 128K output) with support for...
Pricing
Per million tokens · OpenRouter
Input$2.50
Output$15.00
Blended (3:1)$5.63
Specification
openai/gpt-5.4
Context window1.1M
Max output128K
ReleasedMar 2026
LicenceProprietary
Modalitiesin:text, in:image, in:file, out:text
Benchmarks
Independently run — not vendor self-reports. Every score links its source.
| Benchmark | Score | Config | Run | Source |
|---|---|---|---|---|
| GPQA Diamond | 93.3% ±1.8 | xhigh effort | 2026-03-06 | Independently run · Epoch AI |
| SWE-bench Verified | 76.9% ±1.9 | high effort | 2026-03-06 | Independently run · Epoch AI |
| FrontierMath T1–3 | 78.6% ±2.4 | xhigh effort | 2026-06-11 | Independently run · Epoch AI |
| FrontierMath T4 | 49.0% ±8.0 | xhigh effort | 2026-06-11 | Independently run · Epoch AI |
| SimpleQA Verified | 44.8% ±1.6 | xhigh effort | 2026-03-06 | Independently run · Epoch AI |
| OTIS Mock AIME | 97.8% ±2.2 | high effort | 2026-07-15 | Independently run · Epoch AI |
Scores published by Epoch AI, 'AI Benchmarking Hub', used under CC BY 4.0.
Cheaper models that score at least as well on GPQA Diamond
Strictly lower blended price AND an equal-or-higher GPQA Diamond score. One benchmark is one dimension — a model that wins here may still be weaker at your specific task.
| Model | Provider | Blended $/1M | GPQA Diamond | |
|---|---|---|---|---|
| GPT-5.6 Terra | OpenAI | $2.25 (−60%) | 93.3% | Compare → |
| Gemini 3.6 Flash | $3.00 (−47%) | 94.1% | Compare → | |
| Grok 4.5 | xAI | $3.00 (−47%) | 93.4% | Compare → |
| Gemini 3.1 Pro Preview | $4.50 (−20%) | 94.4% | Compare → |
Where to run GPT-5.4
6 options across 3 providers · 4.0× between cheapest and dearest. Blended at 3:1. ⚠ marks an option serving less context than the best available, or re-pricing above a prompt-length threshold. Bracketed labels are the provider's own service tier or region.
| Host | Blended $/1M | In / Out | Context | Uptime 24h |
|---|---|---|---|---|
| OpenAI (flex) ⚠ re-prices above 272K prompt tokens ($2.50/$11.25) |
$2.81 | $1.25 / $7.50 | 1.1M | 97.1% |
| OpenAI ⚠ re-prices above 272K prompt tokens ($5.00/$22.50) |
$5.63 | $2.50 / $15.00 | 1.1M | 97.1% |
| Azure ⚠ re-prices above 272K prompt tokens ($5.00/$22.50) |
$5.63 | $2.50 / $15.00 | 1.1M | 100.0% |
| Azure (eu) | $6.19 | $2.75 / $16.50 | 1.1M | 100.0% |
| Amazon Bedrock (us-east-1) ⚠ re-prices above 272K prompt tokens ($5.50/$24.75) |
$6.19 | $2.75 / $16.50 | 1.1M | — |
| OpenAI (priority) | $11.25 | $5.00 / $30.00 | 1.1M | 97.1% |
Host prices and uptime from OpenRouter, synced 2026-08-08.
Compare GPT-5.4
FAQ
How much does GPT-5.4 cost?
GPT-5.4 costs $2.50 per million input tokens and $15.00 per million output tokens on OpenRouter. At a 3:1 input:output mix that blends to $5.63 per million tokens.
What is GPT-5.4's context window?
GPT-5.4 accepts up to 1,050,000 tokens of context, and can generate up to 128,000 output tokens.
How good is GPT-5.4 on benchmarks?
GPT-5.4 scores 93.3% on GPQA Diamond, independently run and published by Epoch AI. 6 benchmark results are listed on this page, each with its run date and source.
Is there a cheaper model as good as GPT-5.4?
Yes — 4 models in our catalogue cost less per token than GPT-5.4 and score at least as high on the same benchmark. The cheapest is GPT-5.6 Terra at $2.25 per million tokens blended.