Vercel launches GPT 5.3 Codex on AI Gateway; 25% faster coding model now available
GPT 5.3 Codex Now Available on AI Gateway
Vercel has made GPT 5.3 Codex available through its AI Gateway platform. This new model consolidates the coding strengths of GPT-5.2-Codex with the reasoning depth of GPT-5.2, delivering a 25% performance improvement while reducing token consumption.
Key Capabilities
The model is purpose-built for long-running agentic work, handling:
- Research and analysis across codebases
- Multi-step execution spanning the full software development lifecycle
- Debugging, deployment, monitoring, and product documentation
- Data analysis and transformation tasks
- Web development with better handling of underspecified prompts and production-ready defaults
A standout feature is the ability to steer the model mid-task without losing context, enabling more dynamic and iterative workflows.
Integration & Usage
To use the model, set the model identifier to openai/gpt-5.3-codex in the Vercel AI SDK:
import { streamText } from 'ai';
const result = streamText({
model: 'openai/gpt-5.3-codex',
prompt: `Research our current API architecture, identify performance
bottlenecks, refactor the slow endpoints, add monitoring,
and deploy the changes to staging.`,
});
AI Gateway Benefits
Beyond the new model, developers gain access to AI Gateway's unified infrastructure:
- Usage tracking and cost monitoring across multiple providers
- Built-in observability and performance analytics
- Bring Your Own Key (BYOK) support for flexible provider management
- Intelligent routing, retries, and failover for enterprise-grade reliability
- Model leaderboards to benchmark performance across options
Visit the AI Gateway model playground to get started.