What is the context window of GPT-4.5?

GPT-4.5 has a context window of 8K tokens.

When was GPT-4.5 released?

GPT-4.5 was released on 2026-03-12.

Which providers offer GPT-4.5?

GPT-4.5 is available from 2 providers: OpenAI API, Azure OpenAI.

What benchmarks has GPT-4.5 been tested on?

GPT-4.5 has been evaluated on 1 benchmark, including MMLU PRO.

GPT-4.5

Name: GPT-4.5
Author: OpenAI

gpt-4-5

DeprecatedResearched 8d ago

Last refreshed 2026-05-10. Next refresh: weekly.

ProprietaryMultimodalCodingAgentsVisionClassificationJSON / Tool use

GPT-4.5 is a legacy integration reference; evaluate GPT-4.1 before starting new work.
Decision context: Classification task fit, 2 tracked provider routes, and research from 2026-05-10.

Use it for

Teams maintaining an existing integration
Workloads that can use a 8K context window
Buyers comparing 2 tracked provider routes

Do not use it for

New production launches

Cheapest output

Azure OpenAI per 1M tokens

Provider routes

Tracked API hosts

Quality / dollar

Unknown

No output-token price in the ladder

Freshness

2026-05-10

Researched 8d ago

fresh

This API model is marked deprecated. Use GPT-4.1 as the replacement candidate before sending new production traffic.

Top use-case fit

Coding

Included by capability and metadata signals in the decision map.

Agents

Included by capability and metadata signals in the decision map.

Vision

Included by capability and metadata signals in the decision map.

Provider price ladder

Provider	Input / 1M	Output / 1M	Route
Azure OpenAI	-	-	ServerlessPartial
OpenAI API	-	-	ServerlessPartial

Benchmark peer barsfor Classification

MMLU PRORank 17 of 64

90.1

89.1

88.9

88.6

86.1

Migration checks

No linked migration route is available for this model yet.

About

Improved GPT-4 variant released March 2026. Enhanced performance over GPT-4o with better reasoning and code generation.

GPT-4.5 has a 8K-token context window.

Capabilities

VisionMultimodalFunction CallingTool UseStructured OutputsCode Execution

Benchmark Scores(1)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.