OpenAI Releases GPT-5.4 Image 2 with 272K Context Window and Image Generation

OpenAI has released GPT-5.4 Image 2, a multimodal model that combines the company's GPT-5.4 reasoning capabilities with image generation. The model is available via OpenRouter's API under the identifier openai/gpt-5.4-image-2.

Technical Specifications

GPT-5.4 Image 2 features a 272,000 token context window and supports text, image, and file inputs with text and image outputs. OpenRouter lists pricing at $8.00 per million input tokens and $15.00 per million output tokens.

According to OpenRouter, the model "combines OpenAI's GPT-5.4 model with state-of-the-art image generation capabilities from GPT Image 2." The system is designed for what OpenRouter describes as "rich multimodal workflows," enabling users to move between reasoning, coding, and image generation tasks within the same context.

Availability and Access

The model is currently available exclusively through OpenRouter's API. OpenAI has not announced direct API access through its own platform, and no benchmark scores or parameter counts have been disclosed.

OpenRouter positions the model as suitable for workflows that require both analytical reasoning and visual content generation in a single session, leveraging the extended context window to maintain coherence across complex multimodal tasks.

What This Means

GPT-5.4 Image 2 represents OpenAI's continued expansion into multimodal AI, though the exclusive availability through OpenRouter raises questions about the model's official status and whether this is a full public release or a limited partnership rollout. The 272K context window is competitive with other frontier models, but without benchmark data or direct comparison to GPT-4 Vision or other multimodal systems, it's difficult to assess the model's capabilities independently. The pricing sits in the mid-range for multimodal models, making it accessible for production use cases that require both language understanding and image generation.