When was gpt-5.4-pro-2026-03-05 released, and is it still maintained?

GPT-5.4 Pro was released on 2026-03-05. Its catalog status is stable. Check the model page for deprecation notices before you build a long-lived integration on it.

How should you evaluate gpt-5.4-pro-2026-03-05 before production?

Use representative tasks and compare quality, latency, token use, and relevant tool behavior with your current model. Reserve gpt-5.4-pro-2026-03-05 for evaluated difficult or high-value cases, and compare it with a measured baseline before setting routing rules.

Does gpt-5.4-pro-2026-03-05 charge more for long inputs?

Yes — gpt-5.4-pro-2026-03-05 uses tiered pricing for long-context requests. Inputs exceeding a token threshold are billed at a higher per-token rate. Check the model page for the exact threshold and long-context pricing tier applicable to gpt-5.4-pro-2026-03-05.

What controls should you set before using gpt-5.4-pro-2026-03-05 in production?

Set task boundaries, review requirements, failure handling, and monitoring that match the consequences of the workload. Do not rely on an unevaluated model response for irreversible actions.

How should you begin a pilot with gpt-5.4-pro-2026-03-05?

Reserve gpt-5.4-pro-2026-03-05 for evaluated difficult or high-value cases, and compare it with a measured baseline before setting routing rules.

What should you evaluate before moving production traffic to gpt-5.4-pro-2026-03-05?

Before moving production traffic to gpt-5.4-pro-2026-03-05, verify pricing, rate limits, and output quality on your specific tasks. Run a pilot with a representative sample of your workload, check latency at your expected concurrency, and confirm that the model handles your edge cases reliably.

What is the max output length for gpt-5.4-pro-2026-03-05?

gpt-5.4-pro-2026-03-05 supports up to 128000 output tokens per request. Plan your prompts to stay within this limit, and use the max_tokens or max_completion_tokens parameter to control output length.

Model detail

GPT-5.4 Pro 20260305

Provided by OpenAI

Pay-as-you-go Dynamic pricing 2 tiers

GPT-5.4 Pro is OpenAI's most advanced model, building on GPT-5.4's unified architecture with enhanced reasoning capabilities for complex, high-stakes tasks. It features a 1M+ token context window (922K input, 128K output) with support for text and image inputs. Optimized for step-by-step reasoning, instruction following, and accuracy, GPT-5.4 Pro excels at agentic coding, long-context workflows, and multi-step problem solving.

Model specs

Model ID: gpt-5.4-pro-2026-03-05

Context length: 1.05M
Max output: 128K
I/O modalities
Released: 2026-03

API endpoints

One MixRoute gateway, OpenAI-compatible

OpenAI-compatible /v1/chat/completions POST

Tiered pricing

Unit: /1M Tokens

Tier	Input /1M Tokens	Output /1M Tokens
standard Length ≤ 272K	$30.0000	$180.0000
long_context Length > 272K	$60.0000	$270.0000

Selection summary

Quickly judge whether it fits your workload

What is this model good for?

Enhanced reasoning for complex high-stakes tasks

Use gpt-5.4-pro-2026-03-05 for agentic coding, long-context workflows, and multi-step problem solving. It builds on GPT-5.4 with enhanced reasoning for complex tasks.

What should you check before using it?

Confirm the snapshot date matches your stability needs

This is the 2026-03-05 snapshot of GPT-5.4 Pro. Verify that this version provides the stability you need for your most demanding reasoning tasks. Knowledge cutoff: 2025-08-31.

Why use it through MixRoute?

GPT-5.4 Pro runs tools and structured output here

Tool calling and structured output run on that route as well. MixRoute serves GPT-5.4 Pro on the OpenAI-compatible endpoint /v1/chat/completions; point your existing client at https://api.mixroute.ai/v1 with the model ID gpt-5.4-pro-2026-03-05 and re-run your integration tests.

Capabilities & compatibility

Unconfirmed capabilities stay "Unknown," no guesses

Tool calling

GPT-5.4 Pro supports function and tool calls.

EX Send a tools definition; the response returns tool_calls to execute and feed results back.

Streaming

GPT-5.4 Pro supports streamed responses.

EX stream: true streams tokens incrementally, so chat UIs render as they arrive.

Prompt caching

GPT-5.4 Pro supports prompt caching under supported routing conditions.

EX Keep the system prompt and tool definitions at the start of the request; repeated prefixes are automatically billed at the cache-read rate.

Vision input

GPT-5.4 Pro can process image inputs alongside text.

EX Pass a base64-encoded image in a image_url content part alongside text messages.

OpenAI SDK

Call GPT-5.4 Pro through MixRoute's OpenAI-compatible endpoint.

EX Change base_url to https://api.mixroute.ai/v1; everything else follows the OpenAI style.

Token cost estimator

Live estimate from this page's pricing, not an actual bill

Input tokens Max per request 1.05M

Output tokens Max per request 128K

Requests / month

Cache hit rate –

This model has no cache-read price; caching is not counted

FAQ

GPT-5.4 Pro is billed per use and the record publishes no free-tier allowance. The current published rates are input $30.00 per 1M tokens and output $180.00 per 1M tokens. Trial credit or promotional quota is tied to your account rather than to the model, so check the console before planning around a free tier.

Get started

01

Create an API key

Generate a key in the Console; one key calls every model.
02

Set the base URL

Fully OpenAI SDK-compatible; just swap the URL to connect.

https://api.mixroute.ai/v1
03

Specify the model ID

In the request's model field, enter:

gpt-5.4-pro-2026-03-05

Create an API key

New users get $5 in free credits

Test this model and alternate routes with the same request format

Start from a real workload, then let quality, total cost, and failure conditions decide whether to send production traffic.

Create an API key View integration docs