Model
GPT-5.4 Pro
More exact GPT-5.4 tier for demanding professional reasoning and agent tasks
12 providersreleased knowledge cutoff 2025-08-31closed weights
$27
best input /M · Poe
$160
best output /M
1.05M
context window
128K
max output
Compare providers (sorted by listed input price)
| Provider | Input /M | Output /M | Cache read /M | Cache write /M | Context | Capabilities | Status | Updated |
|---|---|---|---|---|---|---|---|---|
| Poe openai/gpt-5.4-pro | $27 | $160 | — | — | 1.05M | reasoningtoolsno structuredvision | Mar 5, 2026 | |
| Impossibl openai/gpt-5.4-pro | $30 | $180 | $3 | — | 1.05M | reasoningtoolsno structuredvision | Mar 5, 2026 | |
| OpenRouter openai/gpt-5.4-pro | $30 | $180 | — | — | 1.05M | reasoningtoolsstructuredvision | Mar 5, 2026 | |
| OpenAI gpt-5.4-pro | $30 | $180 | — | — | 1.05M | reasoningtoolsno structuredvision | Mar 5, 2026 | |
| LLM Gateway openai/gpt-5.4-pro | $30 | $180 | — | — | 1.05M | reasoningtoolsno structuredvision | Mar 5, 2026 | |
| Cloudflare AI Gateway openai/gpt-5.4-pro | $30 | $180 | — | — | 1.05M | reasoningtoolsno structuredvision | Mar 5, 2026 | |
| Ofox openai/gpt-5.4-pro | $30 | $180 | — | — | 1.05M | reasoningtoolsno structuredvision | Mar 5, 2026 | |
| Vercel AI Gateway openai/gpt-5.4-pro | $30 | $180 | — | — | 1.05M | reasoningtoolsno structuredvision | Mar 5, 2026 | |
| Eden AI openai/gpt-5.4-pro | $30 | $180 | $3 | — | 1.05M | reasoningtoolsstructuredvision | Mar 5, 2026 | |
| Kilo Gateway openai/gpt-5.4-pro | $30 | $180 | — | — | 1.05M | reasoningtoolsstructuredvision | Mar 5, 2026 | |
| ZenMux openai/gpt-5.4-pro | $45 | $225 | — | — | 1.05M | reasoningtoolsno structuredvision | Mar 20, 2026 | |
| OrcaRouter openai/gpt-5.4-pro | $60 | $270 | — | — | 1.05M | reasoningtoolsno structuredvision | Mar 5, 2026 |
Price badge
https://models.sutraworks.ai/badge/openai/gpt-5.4-pro.svgLive SVG, regenerated on every hourly sync — paste it into a README to always show the current cheapest listed price. Updates when prices change; no build step needed on your side.
Benchmarks
- GPQA Diamond ↗(accuracy)report ↗94.4
- Humanity's Last Exam ↗(accuracy)report ↗42.7
- Humanity's Last Exam ↗(accuracy)report ↗58.7
- BrowseComp ↗(accuracy)report ↗89.3
- GDPval(wins or ties)report ↗82
- FrontierMath ↗(accuracy)report ↗50
- FrontierMath ↗(accuracy)report ↗38
- ARC-AGI-1 ↗(accuracy)report ↗94.5
- ARC-AGI-2 ↗(accuracy)report ↗83.3
- FinanceAgent(accuracy)report ↗61.5
- GeneBench(accuracy)report ↗25.6
Prices are per million tokens (USD). “—” means the provider does not publicly list a price for this model. Data from models.dev; verify with the provider before purchasing.