Pull down to go back
OpenAI's new image generator can now search the web while creating pictures

OpenAI's new image generator can now search the web while creating pictures

OpenAI 新版圖片生成器現在會上網查資料,幫你生成更精準的圖片

OpenAI just rolled out ChatGPT Images 2.0, and it's getting a serious upgrade. The new version can now browse the web in real-time to help generate images—so instead of just guessing what you want, it can actually look things up to make your pictures more accurate and detailed. The update includes better instruction-following, sharper text generation, and the ability to preserve specific details you care about. It's powered by OpenAI's fresh GPT Image 2 model with "thinking capabilities," and it's available to ChatGPT Plus, Pro, Business, and Enterprise users. Basically, your AI image generator just got smarter about understanding what you're actually asking for.

Tech Blogger Take

Your AI image generator just learned to Google. This changes everything.

OpenAI dropped ChatGPT Images 2.0 and it's not just another incremental update — this thing can now browse the web while creating your pictures. Think about what that actually means: no more getting a generic 'iPhone-looking device' when you ask for the iPhone 15 Pro. No more 'close enough' logos when you need the actual Nike swoosh. The AI can literally look up what you're talking about and get it right the first time. The new GPT Image 2 model comes with 'thinking capabilities' (whatever that means) and better instruction-following, plus it can preserve specific details you care about. It's rolling out to Plus, Pro, Business, and Enterprise users right now. We've gone from 'AI that makes pretty pictures' to 'AI that makes the RIGHT pictures.'

VerdictIf you're still using Midjourney for client work, log into ChatGPT Plus right now and watch your revision count drop to zero.
8/10

Action

版本更新
1Upgrade to ChatGPT Plus ($20/month) to access Images 2.0
2Try a prompt with specific, searchable references like 'iPhone 15 Pro in Titanium Blue on a marble desk'
3Compare results with your old image generators — notice the accuracy difference
Before

Spending 20 minutes describing generic 'smartphone-like device' and getting close-but-wrong results

After

Asking for 'iPhone 15 Pro in Natural Titanium' and getting pixel-perfect product imagery in 30 seconds

AI Analysis

Creative & Marketing

high
Action Required

Test ChatGPT Images 2.0 with your most challenging briefs this week — the ones where you usually need three rounds of revisions

Key Insight

This isn't just better image generation — it's the first AI that can fact-check itself while creating, meaning no more 'close enough' results when you need specific logos, products, or current events

Why It Matters

Your clients will notice when your mockups suddenly have the right font on that billboard or the correct product packaging in lifestyle shots

Job Impact Analysis

Graphic Designer

Role Shift
Why It Impacts

Real-time web search means AI can now reference current brand guidelines, specific products, and accurate visual details instead of hallucinating generic versions

How to Adapt

Upgrade to ChatGPT Plus today and test it on your next client project that requires brand-accurate imagery

Content Creator

Opportunity
Why It Impacts

Better instruction-following and web-aware generation means fewer iterations to get the exact scene, product, or reference you need for your content

How to Adapt

Start incorporating specific, searchable references in your image prompts — mention exact products, locations, or current events

Glossary

Real-time web search(即時網路搜尋)
The ability for AI to browse the internet while generating content, like how ChatGPT Images 2.0 can look up current information about products or brands to make accurate images instead of guessing.
Instruction-following(指令遵循)
How well an AI understands and executes specific requests — the upgrade means ChatGPT Images 2.0 is better at capturing the exact details you ask for rather than giving you something 'close enough.'
Thinking capabilities(思考能力)
OpenAI's vague term for the new GPT Image 2 model's enhanced reasoning — presumably meaning it can better understand context and make smarter decisions about what to include in your images.