Agentic Engineering Power
Cursor-native training enables superior terminal control and multi-step bug resolution in large codebases.
grok-4.5 / image-to-text
Grok-4.5 is a frontier reasoning model by SpaceXAI optimized for agentic software engineering. It features a 500k context window and V9 architecture, delivering high-speed technical performance with 4x better token efficiency than rivals.
$ 1.2
$ 2
image
$ 3.6
$ 6
text
Grok-4.5 Core Technical Features
The technical pillars that make Grok-4.5 a leader in agentic AI performance.
500k Context Window
Massive context support allows for the ingestion of entire repositories or lengthy legal and financial documents.
High-Throughput Serving
Reliable speeds of 80–93 tokens per second ensure your real-time applications remain responsive under load.
V9 Architecture Scale
A 1.5 trillion parameter foundation specifically tuned for STEM, engineering, and complex technical datasets.
How to Get a grok-4.5 API Key
Getting a grok-4.5 API key takes four steps and a few minutes. Create a free GPTProto account, add credits, generate your key, and make your first call — at $1.2 / $3.6 it's a cheaper grok-4.5 API key than going direct, and one key works across every model on the platform. Full grok-4.5 Documentation is in the docs.
Sign up
Create your free GPT Proto account to begin. You can set up an organization for your team at any time.
Top up
Your balance can be used across all models on the platform, including grok-4.5, giving you the flexibility to experiment and scale as needed.
Generate your API key
In your dashboard, create an API key — you'll need it to authenticate when making requests to grok-4.5.
Make your first API call
Use your API key with our sample code to send a request to grok-4.5 via GPT Proto and see instant AI-powered results.
Grok-4.5 Frequently Asked Questions
How does Grok-4.5 handle complex coding tasks?
Grok-4.5 is optimized for terminal use and multi-step issue resolution. It scored 83.3% on Terminal Bench 2.1, proving its excellence in managing file systems and fixing bugs in large repositories.
What are the token costs for Grok-4.5?
Our pricing is highly competitive at $2.00 per 1M input tokens and $6.00 per 1M output tokens. You can also save 75% on cache hits with prompt caching and 50% via the Batch API.
Does Grok-4.5 support multimodal inputs?
Yes, Grok-4.5 provides native support for images and audio. This allows developers to build multimodal vision and voice-driven workflows with ultra-low latency.
Is my data used to train the Grok-4.5 model?
No. Our enterprise agreement with SpaceXAI ensures that no data sent through our API endpoints is used for training foundation models, maintaining strict privacy for your technical work.
How do I migrate to Grok-4.5 from GPT-5.5?
Migration is seamless as our API is 100% OpenAI compatible. Simply update your base URL and change the model parameter to grok-4.5 to start benefiting from higher token efficiency.
What is the reasoning_effort parameter?
This native parameter allows you to scale the internal Chain of Thought. Set it to 'high' for difficult logic tasks or 'low' for faster, more direct responses depending on your needs.