GPT-5.4 is priced at $2.50/1M input and $15/1M output tokens while GPT-5.4 Pro is $30/1M input and $180/1M output tokens, more than GPT-5.2 and GPT-5.2 Pro
and You Can Interrupt It When It Goes Off TrackOpenAI Developers:Pricing — Text tokens Prices per 1M tokens. Batch Flex Standard Priority Model Input …Forums:BeauHD /Slashdot:OpenAI Releases New ChatGPT Model For Working In Excel and Google Sheets
Context & Ripple Effects
OpenAI is pairing GPT-5.4’s higher API rates with a product push into spreadsheet work and an interruptible workflow. The same release also expands the API’s tool-calling capability and offers up to a 1M-token context window, making the pricing consequential for developers building longer-running tasks.
This is an early step in a pricing arc: later coverage says GPT-5.5 doubled GPT-5.4’s standard token rates, while subsequent GPT-5.6 tiers reintroduced lower-priced options. The key question is therefore not token price alone, but whether higher-cost models deliver enough additional task value.
First-order effects
- Developers using GPT-5.4 or GPT-5.4 Pro face higher per-token spend than with GPT-5.2 equivalents, especially on output-heavy workloads.
- OpenAI creates a clearer premium tier for customers that value stronger spreadsheet-oriented work, tool calling, and long-context tasks over the lowest API bill.
Second-order effects
- Application teams will need to tighten model routing, output controls, and workload measurement, because long-context and agentic workflows can turn higher output rates into materially larger task costs.
- Competing model providers have more room to position lower-priced models for routine workloads, while OpenAI’s customers must test whether GPT-5.4 reduces retries or human intervention enough to justify its price.
Third-order effects
- The market is moving from a simple race to lower token prices toward segmented model pricing, where premium reasoning or agent-capable tiers coexist with cheaper alternatives for less demanding work.
- As models are embedded in business workflows, procurement is likely to focus increasingly on cost per completed task rather than published token rates; that favors providers and buyers able to measure quality, retries, and human oversight together.
The trend: AI API pricing is becoming more tiered and workload-specific as providers monetize higher-capability models while customers optimize for cost per useful task.