View more inline images in answers
TL;DR
Free users globally will now see more inline web images for topics like people, places, and products. This update applies to responses generated using the 5.5-Instant model on web and iOS.
What changed
Free users globally now receive more inline web images in ChatGPT answers for topics involving people, places, and products. The change applies only to responses from the 5.5-Instant model on web and iOS. It began rolling out on May 12, 2026.
Users can click any image to open the original size and view source attribution. Images appear automatically when the model decides visuals add clarity. No extra settings or plan upgrades are required.
Why it matters
This move improves the free tier experience for quick visual lookups without forcing users into paid plans. It raises the baseline quality of everyday answers and may increase time spent in ChatGPT by free accounts.
The bet appears to be that better free responses will drive habit formation and later conversions. It also pressures competitors who still limit images behind paywalls.
How to use it
Open ChatGPT on web or iOS while signed in with a free account. Ask questions about well-known people, places, or products using the default 5.5-Instant model.
Images surface automatically in relevant answers. Click any image for the full view and source link. The feature works in standard chats and does not require temporary chat mode.
Watch for
Confirm the bet if free users report fewer follow-up questions and higher satisfaction on visual topics. It breaks if images frequently mismatch context or load slowly.
Expect OpenAI to expand the same image treatment to additional models or add controls for image frequency next.
Who this matters for
- Vibe Builders: Use the 5.5-Instant model on iOS to quickly source and verify visual references for mood boards.
- Basic Users: Ask for travel or product info on the free tier to see inline images with direct source links.
Harsh’s take
OpenAI is aggressively raising the floor for free users to maintain market dominance. By integrating inline web images into the 5.5-Instant model, they are turning ChatGPT into a visual search engine that rivals traditional browsers. This move targets habit formation: if a user can get the text and the visual proof in one interface, they stop bouncing to Google Images.
For operators, this is a reminder that the free tier is a product, not just a lead magnet. The inclusion of source attribution is a smart play to mitigate hallucination concerns while keeping users within the ecosystem. It puts immediate pressure on competitors who still treat basic visual retrieval as a premium feature.
Expect this to increase session depth for casual users who previously only used the tool for text-heavy tasks.
by Harsh Desai
About ChatGPT
View the full ChatGPT page →All ChatGPT updatesGo deeper
More AI news
- FeatureHermes Agent verifies work with completion contracts and evidence ledgers
Hermes Agent records verification evidence for coding tasks. The /goal command uses completion contracts to judge success against test runs rather than model assertions.
- FeatureCursor adds cloud agent management to the Agents window
Cursor sets up cloud development environments in under 10 minutes, spins up isolated cloud subagents using /in-cloud, and hands off sessions between local and cloud.
- FeatureCursor introduces /automate skill for automating repetitive tasks
Cursor's new /automate skill creates automations from plain language. Workflows trigger via Slack emojis or GitHub events while cloud agents access virtual computers.