OpenAI Patches GPT-5.4 Image Encoder Bug That May Improve Vision Quality

Mar 15, 2026 · Updated Apr 25, 2026

GPT-5.4, OpenAI's frontier model for professional work released earlier this month, received an image encoder fix on March 13, 2026 — a bug patch addressing issues with input_image inputs in both the v1/responses and v1/chat/completions APIs. The fix updates how the model processes image data at the encoder level, and OpenAI classifies this as a bug fix, not a feature change.

If you're building with GPT-5.4 and relying on image understanding, some of those use cases may now see improved quality in outputs. No changes are required from developers.

Try your existing image understanding workflows against GPT-5.4 now and compare outputs — improvements will vary by use case.

View the full update on developers.openai.com

OpenAI Developers

@OpenAIDevsMar 13

We updated our image encoder to fix a small bug for image inputs GPT-5.4. Some image understanding use cases may now see improved quality. No action needed. https://t.co/OUvMWsRRtm

View on X