TokenTally

OpenAI frontier models

GPT-5.4 cost planner

At $2.50/M input and $15/M output for short runs, GPT-5.4 scales to 1.05M tokens when you need it.

Last pricing check: Mar 13, 2026

$2.50 per 1M prompt tokens$15.00 per 1M completion tokens1,050,000 token context

Why teams choose this model

Executive copilots with long briefs
Enterprise-grade RAG agents
Model-to-model evaluation
Open main calculator

Scenario planning

Realistic cost examples

Numbers use GPT-5.4 pricing

Board memo assistant

Leadership teams feed full project dossiers into GPT-5.4 for ready-to-send updates.

Per request

$0.031

Per month

$130.20

Tokens sent

18,480,000

2800 prompt tokens1600 completion tokens4,200 requests/mo

Autonomous planner

Agent chains plan multi-day launches with structured tool calls and long reasoning.

Per request

$0.041

Per month

$246.00

Tokens sent

32,400,000

3200 prompt tokens2200 completion tokens6,000 requests/mo

Quant research bench

Quant teams spin GPT-5.4 to audit trading playbooks with citations and confidence scoring.

Per request

$0.0255

Per month

$229.50

Tokens sent

28,800,000

1800 prompt tokens1400 completion tokens9,000 requests/mo

Compare with

OpenAI

GPT-5 Pricing Guide

Updated Mar 13, 2026

OpenAI

GPT-5.1 Pricing Guide

Updated Mar 13, 2026

OpenAI

GPT-5.2 Pricing Guide

Updated Mar 13, 2026

FAQs

When do long-context rates kick in?

OpenAI charges double input / 1.5x output once a GPT-5.4 prompt exceeds 272K tokens. TokenTally keeps estimates under that line by default.

Is cache pricing available?

$0.25/M tokens applies when the prompt prefix is cached via Responses/Assistants APIs. We expose that as the cache-hit number in the calculator.

How do I treat reasoning tokens?

Reasoning tokens count toward output spend even though they don’t appear in responses. Budget with a headroom factor if you enable extra thinking time.

Pricing sources