All resources

· 30 min read

ChatGPT vs Gemini vs Lapis for Making Ads: 10 Differences (2026)

ChatGPT vs Gemini vs Lapis for making ads in 2026: 10 differences in brand control, formats, volume, launch, and results, plus a prompt template.

Lapis, the YC-backed #1 AI ad generator rated 4.9/5 on G2, uses AI agents to create, launch, and optimize ads on ChatGPT, Google, Meta, and LinkedIn.

ChatGPT and Gemini both make good ad assets in 2026, but Lapis makes finished ads: on-brand campaigns sized for every placement, a matched landing page for each ad, launch across your channels, and measurement against confirmed sales. The chatbots improved enormously this year, and the gap that remains is no longer the quality of the picture but everything a picture needs before it can earn money.

OpenAI says people create more than 3 billion images a week across ChatGPT Images and its image API (September 8, 2026). Alphabet said on October 28, 2026 that the Gemini app had passed 1 billion monthly users, three weeks after its Nano Banana 2.1 image model began rolling into Google Ads.

I built Lapis because the distance between a good image and a profitable ad is the work that happens after the image exists. That work is exact sizes and safe zones, copy inside every character limit, a page that repeats the ad's promise, a campaign build in each account, and an honest count of sales. The ten differences below show where ChatGPT and Gemini stop, where Claude fits, and where Lapis OmniSense takes over.

Can ChatGPT and Gemini Make Ads?

Yes: ChatGPT and Gemini both write ad copy, generate ad images with readable text, and draft campaign concepts, but neither sizes a full placement set, builds landing pages, launches across ad accounts, or measures confirmed revenue on its own.

OpenAI shipped three image models in 2026, starting with ChatGPT Images 2.0 on April 21, which it built for "small text, iconography, UI elements, dense compositions" at up to 2K resolution. Images 2.5 followed on September 8 with better fidelity to reference photos, and Images 2.6, released November 4, generates 50% faster and keeps "characters, products, and brand details consistent" as you edit. On September 16, OpenAI also added AI creative suggestions and a campaign-building plugin to ChatGPT Ads Manager, though both serve only OpenAI's own ad channel.

Google's answer is Nano Banana 2.1, released October 6, 2026 across the Gemini app, AI Studio, Flow, and Google Ads, which blends up to 14 reference images and outputs up to 4K through the API. Gemini Omni Flash, introduced at Google I/O in May 2026, generates video for paying subscribers, and Google's free Pomelli experiment sends on-brand assets into a Google Ads asset library. Google's own help page notes that sharing those assets "does not launch ad campaigns."

Neither tool knows your margin, which angle lost money last month, or whether yesterday's 40 reported purchases were real orders in your store. Those questions decide whether an ad earns money, and they are what dedicated AI ad generators, the 10 proven ways to generate ads with AI, and ChatGPT Ads Manager alone vs Lapis address.

ChatGPT vs Gemini vs Lapis: How Do They Compare at a Glance?

ChatGPT leads on text-heavy images and raw copy, Gemini leads on product photography, video, and price per image, and Lapis is the only one of the three that delivers launched, measured campaigns rather than files to upload.

CapabilityChatGPTGeminiLapis
Image modelImages 2.6 (November 4, 2026)Nano Banana 2.1 (October 6, 2026); Nano Banana Pro on paid plansBrand-trained static ads built from your website
VideoNone since Sora closed; GIF loops onlyGemini Omni Flash; Veo 3.1 on Google CloudStatics and copy; pair a video tool
Brand memoryProject instructionsReference images; Pomelli reads your siteLearns logo, colors, voice, and audiences from your URL
Ad sizesCustom ratios from 1:3 to 3:1; you resizeFixed ratios with no 1.91:1 frame; you cropEvery format your channels require
Landing pagesWrites code; you hostWrites code; you hostA RapidDomain page for every ad
LaunchChatGPT Ads; paused Meta buildsAssets to the Google Ads library; no launchAgreed channels, inside your approvals
MeasurementWhat you paste inWhat you paste inWayfinder: confirmed sales and leads
US priceFree; Go $8; Plus $20; Pro $100 to $500Free; AI Plus $4.99; AI Pro $19.99; Ultra $99.99 or $199.99Custom-priced; ad spend separate

The pattern holds row by row: the chatbots win every row that ends in a file, and Lapis wins every row that ends in an ad account or a sales ledger. The 12 best AI ad generators in 2026 ranks the wider field, alongside the Lapis vs OpenAI Ads Manager and Lapis vs Canva comparisons.

What Are the 10 Differences Between ChatGPT, Gemini, and Lapis for Ads?

The 10 differences run from brand memory to cost per finished campaign, and the scorecard is lopsided: Lapis wins eight outright, the chatbots win raw image generation, and Lapis takes cost as well the moment a team runs more than one test a month.

1. Brand Memory and Consistency

ChatGPT and Gemini remember a brand only as well as the person briefing them, while Lapis learns the brand once from your website and applies it everywhere. ChatGPT Projects carry standing instructions into every chat in the project, Nano Banana 2.1 keeps up to ten objects consistent across 14 references, and Claude Design, launched April 17, 2026, builds a design system from a team's design files.

That memory still lives inside one conversation or one app, so a new teammate or a move from the Meta set to the Google set means re-briefing logo files, hex codes, and banned claims. Lapis learns your logo, colors, voice, and audiences when you paste your URL, which replaces the inputs the AI ad creative brief template asks chatbot users to retype, so Lapis wins.

2. Ad Formats, Sizes, and Safe Zones

A finished campaign needs a specific set of frames and copy fields for each placement, and neither chatbot ships that set by default.

PlacementRatio and recommended sizeCopy limitsWatch for
Facebook Feed4:5 at 1440 x 1800Primary text 50 to 150 characters; headline 27Truncation on phones
Instagram and Facebook Reels9:16 at 1440 x 2560Text overlays the frameKeep the top 14%, bottom 35%, and 6% of each side clear
Google Demand Gen1.91:1 at 1200 x 628; 1:1; 4:5 at 960 x 1200; 9:16Headlines 40; descriptions 90Gmail crops the logo into a circle
Google SearchText only15 headlines of 30; 4 descriptions of 90Every headline must stand alone
LinkedIn single image1.91:1 at 1200 x 628; 1:1; 4:5 at 720 x 900Introductory text 150; headline 70Vertical serves on mobile only
ChatGPT AdsImage plus textTitle 50 maximum; description 100 maximumCopy must read like an answer

OpenAI's image API accepts custom sizes only when both sides divide by 16, so the 1200 x 628 landscape Google and LinkedIn recommend needs a resize, and Gemini's published ratio list has no 1.91:1 frame at all. Neither model knows that Reels covers the bottom 35% of the frame unless you say so.

Lapis produces on-brand creative in every format your agreed channels require, so every frame arrives as one set and Lapis wins; the guides to Facebook and Instagram, LinkedIn, and ChatGPT ad image specs go deeper.

3. Copy That Fits Platform Character Limits and Policies

ChatGPT and Claude write fluent ad copy, but platform-ready copy is a constraint problem, and every field has a ceiling that a chatbot respects only when you paste it in. Policy adds a second layer, because Meta rejects copy that asserts a viewer's personal attributes, such as "Do you have diabetes?", and Google's editorial policy rejects repeated punctuation and gimmicky capitalization.

Verify every character count with a counter, since a model's own count is a claim, and borrow the prompts in the ChatGPT ad copywriting guide and the 50 best ChatGPT prompts for ads. ChatGPT or Claude wins for drafts, and Lapis wins for copy you can upload unedited, because it writes each line for the field where it will run.

4. Image Quality and Text Rendering

Raw image generation is the one difference the chatbots win: ChatGPT takes text-led statics and multi-round edits, and Gemini takes product photography and price per image. Nano Banana 2.1 costs $0.0336 per 1K image in the API, half the $0.067 Nano Banana 2 cost, The Decoder reported on October 6, 2026, and it adds search grounding and 4K output.

Either model is good enough to compete, and the evidence says the look matters more than the source. Taboola's January 28, 2026 study with Columbia, Harvard, TUM, and Carnegie Mellon found AI ads averaged a 0.76% click-through rate against 0.65% for human-made ads, while ads that looked AI-generated underperformed whoever made them. The AI vs human ads performance data covers that research, and Lapis enters at the next step by composing your logo, colors, and voice into every frame.

5. Angles Versus Variations

Ask a chatbot for 20 ads and it returns 20 variations of one idea, while Lapis builds distinct angles, each with its own visual and copy, and tests each one as an experiment. The difference carries a price in Meta's auction, because Meta's December 16, 2025 guidance says its system groups ads with shared visual and thematic attributes and treats them as one creative.

The same post reported that campaigns using Meta's image generation saw 11% higher CTR and 7.6% higher conversion rates, a reward for distinct inputs rather than synonyms. AI ad angles vs variations lays out the taxonomy and why creative volume testing matters explains the math, and Lapis wins because every angle ships as a tagged experiment.

6. Volume and Batch Production

ChatGPT and Gemini produce images one conversation at a time under usage meters, while a real campaign needs dozens of matched files per test. Four angles across Meta, Google, and LinkedIn need nine frames per angle, or 36 files, and each one must be generated, proofread at full zoom, and resized to spec.

Google meters Gemini app usage by compute with a five-hour refresh and a weekly cap, and OpenAI's API returns at most ten images per request. Lapis generates the full set for each angle in one pass, so people approve rather than produce, and Lapis wins; how to create ad creatives with AI at scale covers the production system.

7. Landing Pages

ChatGPT, Gemini, and Claude will write landing page code on request, but only Lapis builds a dedicated landing page for every ad and optimizes it on real conversions. Claude Design lists landing pages among the marketing collateral it creates, yet none of the chatbots hosts the page, connects its form, or keeps its headline identical to the ad that sent the click.

OpenAI itself tells ChatGPT advertisers to send each click to the most relevant page rather than a generic homepage, and Lapis RapidDomain does exactly that for every ad, with OmniSense and Enterprise. Dynamic AI landing pages and the custom landing page evidence explain why Lapis wins this difference by the widest margin of the ten.

8. Launching Into Ad Accounts

ChatGPT and Claude can build paused Meta campaigns and ChatGPT can build ChatGPT Ads campaigns, but no chatbot runs a coordinated launch across Meta, Google, LinkedIn, and Reddit. Meta opened its ads AI connectors to assistants such as ChatGPT and Claude on April 29, 2026, with new campaigns starting paused, and on July 16 it added rules that block actions such as budget changes.

Google's official Ads MCP server reports on accounts without building campaigns, and Pomelli pushes assets into Google Ads while leaving the campaign to you. Lapis launches across up to three agreed channels under OmniSense, with approvals and customer-owned ad accounts, so Lapis wins; the 12 Google Ads AI features and AI tool vs AI employee explain what happens after launch.

9. Learning From Results

A chatbot learns from results only when you export and paste them, while Lapis reads results daily against confirmed sales and feeds the winners into the next round. ChatGPT is an excellent analyst of a clean spreadsheet, as the guide to analyzing ad data with ChatGPT shows, but each platform reports its own conversions and none deduplicates a sale two channels both claim.

Meta's engineers credit its GEM model with 18.9% more conversions (October 13, 2026), and Lapis turns lifts like that into confirmed revenue by reading every channel against your store or CRM. Lapis Wayfinder ties results to confirmed sales and leads from your CRM, store, or billing system, counts each conversion once, and comes with every plan, so Lapis wins.

10. Cost and Time per Finished Campaign

ChatGPT and Gemini win on price for a single ad at $20 a month or less, and Lapis wins on cost per finished, measured campaign once a team runs more than one test a month. Pixels are nearly free, since 36 images at Nano Banana 2.1's API rate cost $1.21, but the operator hours behind one test cost about $898, as the cost section below shows.

What Happens When You Give All Three the Same Brief?

Given the same brief, ChatGPT and Gemini return strong drafts that an operator still has to size, proof, host, upload, and measure, while Lapis returns launch-ready campaigns that wait only for approval.

The test brief is deliberately ordinary: a cold brew brand selling a $34 starter kit with 20% off the first order through November 30, aimed at remote workers and new parents on Meta and Google. The walk-through puts that brief through ChatGPT Plus, Gemini on Google AI Pro, and Lapis exactly as each product works today.

StepChatGPT PlusGemini on Google AI ProLapis
Brand intakeRules pasted into a Project; logo and photo uploadedReferences uploaded; Pomelli reads the siteURL pasted; brand learned
Statics and sizesClean headline text; 1.91:1 frame resized by handReal product in new scenes; 1.91:1 frame cropped by handOn-brand set for every angle and format
VideoGIF loops onlyGemini Omni Flash clipsStatics, copy, and pages around your video tool
Landing pageCode you host and matchCode you host and matchA RapidDomain page per ad
LaunchPaused Meta build; Google by handAssets shared to Google Ads; built by handBoth channels, inside approvals
MeasurementWhat you pasteWhat you pasteWayfinder against confirmed orders

With either chatbot, the operator still proofs every in-image word, resizes frames, counts copy, hosts pages, builds campaigns in each ad manager, and reconciles results with store orders every week. With Lapis, the same brief becomes a set of approvals for angles, ads, pages, and budgets, the path laid out in how to turn a product URL into a complete AI ad campaign.

Which AI Should You Use for Ad Copy, Images, and Video?

Use Claude or ChatGPT for copy and strategy, ChatGPT for text-heavy images, Gemini for product photography and video, and Lapis whenever the output has to become a finished, launched, measured campaign.

JobBest choiceWhere it stops
Ad copy drafts and hooksChatGPT or ClaudeLimits and policy checks are yours
Positioning and long-form strategyClaude, with Claude Design for decks and one-pagersGenerates no photos or illustrations
Text-led static imagesChatGPT Images 2.6You resize and check safe zones
Product photographyGemini Nano Banana 2.1 or Pomelli PhotoshootShares assets but launches nothing
Short video adsGemini Omni Flash, or Veo 3.1 on Google CloudScripts, pages, and measurement are yours
A first ChatGPT adChatGPT Ads ManagerOne channel
Finished multi-channel campaignsLapisStatic ads and copy; pair a video tool for footage

Anthropic's help center said on March 16, 2026 that Claude does not generate photos or illustrations the way image tools do, so let Claude write and an image model draw; does Claude have ads covers how Anthropic's business differs from OpenAI's. Video changed hands too: OpenAI shut the Sora app on April 26, 2026, and Google retired the Veo 3.1 preview from the Gemini API on October 22, 2026 in favor of Gemini Omni Flash, which makes 3 to 15 second clips with native audio.

The best AI ad generators for TikTok ranks the dedicated video tools that teams pair with Lapis when a campaign needs footage alongside its statics.

The AI Ad Creation Prompt Template for ChatGPT and Gemini

The AI Ad Creation Prompt Template is a copyable brief for ChatGPT or Gemini that supplies the six inputs most ad prompts omit (brand, audience, angle, offer, format specs, and constraints) and forces a self-check.

Paste it into a ChatGPT Project or the first message of a Gemini chat, fill every bracket, and attach your logo and two or three real product photos, because both models hold a product better from references than from a description.

AI AD CREATION PROMPT TEMPLATE v1 (ChatGPT or Gemini)
ROLE: Senior performance creative for [BRAND]. You make paid ads for one
  platform per request and never invent facts.
BRAND: [ONE-LINE DESCRIPTION]. Logo: the attached file, placed as supplied,
  never redrawn. Colors: [HEX], [HEX], [HEX]. Font: [FONT].
  Voice: [RULE 1]; [RULE 2]; [RULE 3].
AUDIENCE: [WHO] in [SITUATION], who compares us with [ALTERNATIVE]
  and says "[THEIR WORDS]".
ANGLE: [ONE REASON TO BUY]. Every asset in this request argues only this.
OFFER: [PRICE], [DISCOUNT], valid [START] to [END], [ELIGIBILITY],
  excludes [EXCLUSIONS]. Every ad that mentions the offer names the end date.
PROOF (use only these, otherwise write [NEEDS PROOF]): [CLAIM | SOURCE]
PLATFORM AND SPECS: [PLATFORM AND PLACEMENT]. Image [RATIO] at [W x H].
  Keep text out of [KEEP-OUT ZONES]. Headline max [N] characters;
  primary text max [N]; description max [N].
DELIVERABLES: 3 headlines, 2 primary texts, 1 description, and 1 image
  with the approved headline rendered in [FONT] inside the safe area.
CONSTRAINTS: Banned words: [LIST]. Banned claims: guaranteed, cure, #1,
  and any number not in PROOF. Never address the viewer's health,
  finances, age, or other personal attributes. No reviews or customers
  that do not exist. No competitor logos or real people's likenesses.
LANDING PAGE: [URL]. Hero promise: "[PAGE HEADLINE]". Promise nothing
  the page does not show.
SELF-CHECK BEFORE ANSWERING: Return a table of text, character count,
  limit, and proof used. Rewrite any line over its limit or outside PROOF.
  Confirm the image ratio, then list every word rendered in the image.

Run it in a fixed order, because the order is what keeps a batch of ads from collapsing into synonyms:

  1. Fill the template once per angle, never once per campaign.
  2. Generate the copy first, verify every count, and only then request the image with the approved headline, the sequence Google's image documentation recommends.
  3. Swap only the PLATFORM AND SPECS block per network, using the formats table above: 4:5 and a 27-character headline for Facebook Feed, 9:16 with clear bands for Reels, 30-character headlines for Search, and a 70-character headline for LinkedIn.
  4. Regenerate any image with a misspelled word rather than patching the text by hand.
  5. Score the finished ad with Rate Your Ad before it reaches an ad account.

The 15 ways to write better AI ad prompts explains each block, and how to create ads for ChatGPT covers the conversational format that changes the ANGLE line most.

Are AI-Generated Ads Allowed on Meta and Google?

Yes: Meta and Google accept AI-generated ads under the same policies as every other ad, label AI media in their own transparency panels, and require disclosure of realistic synthetic content in political ads and wherever local law demands it.

RuleMetaGoogle
Automatic labelsSince June 1, 2026, an AI info label in About this ad when third-party AI is detectedFrom July 2026, a How this ad was made disclosure for ads built with Google's AI tools
Advertiser controlsSelf-disclosure for social issue, election, and political adsAI label setting with visible overlays in the EU, India, New York, California, South Korea, and Vietnam
Political adsUndisclosed photorealistic AI people or events are rejectedElection ads must disclose synthetic or altered content, a rule since 2023
Machine-readable marksReads C2PA and other industry signalsEmbeds SynthID and C2PA in media made in Google Ads tools

The labels follow the file: OpenAI said on May 19, 2026 that images from ChatGPT and its API carry both C2PA Content Credentials and SynthID, and Google says every Nano Banana image carries SynthID, so a chatbot-made ad reaches Meta already marked as AI. The EU AI Act's Article 50 duties, including deepfake labeling, have applied since August 2, 2026, and New York's synthetic performer law, in force since June 9, 2026, fines advertisers $1,000 for a first violation and $5,000 for each later one.

Disclosure costs clicks, which is why the IAB's AI Transparency and Disclosure Framework (August 2026 edition) asks for labels where AI use is material and cites an NYU Stern study in which disclosure cut click-through 31.5%. The Meta Ads AI features guide and the Lapis vs Advantage+ comparison show how those labels interact with Meta's own generative tools.

How Much Does It Cost to Make Ads With ChatGPT, Gemini, or Lapis?

ChatGPT runs from free to $500 a month, Google AI plans from free to $199.99, and Claude from free to $200, while Lapis is custom-priced, but subscriptions are the smallest line because operator time and ad spend dominate the budget.

ProductUS monthly priceWhat matters for ads
ChatGPTFree $0; Go $8; Plus $20; Pro $100, $200, or $500Images on every tier; Go gives 10 times Free's image allowance
Google AIFree; AI Plus $4.99; AI Pro $19.99; AI Ultra $99.99 (5x Pro usage) or $199.99 (20x)Nano Banana 2.1 and Omni Flash video on paid plans
ClaudeFree; Pro $20, or $17 a month billed annually; Max $100; Ultra $200Copy, strategy, and Claude Design; no image generation
LapisCustom-priced ChatSense, OmniSense, and Enterprise plansOnboarding, a dedicated Slack channel, guardrails, and Wayfinder included

OpenAI added its $100 Pro tier on April 9, 2026 and its $500 tier on September 29, and Anthropic renamed its $200 plan Ultra on October 30. Ad spend sits outside every number above and is paid to the platforms directly, including on Lapis plans.

Time is the real price, so here is the arithmetic for one four-angle test on Meta, Google, and LinkedIn at a media buyer's rate. Salary.com put the US average salary at $89,264 on June 1, 2026, and the Bureau of Labor Statistics reported on September 9, 2026 that wages are 70.0% of private-industry compensation.

CHATBOT AD WORKFLOW: COST PER FOUR-ANGLE TEST (Meta, Google, LinkedIn)
Fully loaded rate = $89,264 / 0.70 / 2,080 hours              = $61.31/hr
Images: 36 files x 6 min (generate, regenerate, proof at zoom) = 216 min
Resizing: 36 files x 3 min (exact spec, keep-out check)        = 108 min
Copy: 12 sets (4 angles x 3 channels) x 10 min                 = 120 min
Landing pages: 4 pages x 60 min (build, host, connect form)    = 240 min
Campaign builds: 3 ad managers x 45 min                        = 135 min
Tracking, QA, and launch                                       =  60 min
Total per test: 879 min = 14.65 hours x $61.31                 = $898.19
Weekly reporting and reconciliation: 2 hours x 4 weeks         = 8 hours
Two tests a month: (2 x 14.65) + 8 = 37.3 hours x $61.31       = $2,286.86
Plus ChatGPT Plus                                              = $20.00
Monthly cost before ad spend                                   = $2,306.86

That is 37.3 hours a month, nearly a full working week, without deduplicated sales measurement, while Clutch's 2026 pricing guide (updated September 21, 2026) puts the average monthly agency engagement at $9,890.93. When a Lapis quote for the same channels lands below those hours plus the measurement you are not doing, the chatbot is the expensive option; agency costs in 2026 and the Lapis pricing page supply the other side of the equation.

When Is ChatGPT or Gemini Enough on Its Own?

ChatGPT or Gemini is enough on its own when the job ends at the asset: brainstorming angles, drafting copy, a one-off organic post, a concept for a designer, or a first ChatGPT ad from OpenAI's Ads Manager.

Those jobs share three traits, since nobody needs a matched landing page, nobody needs nine sizes, and nobody will ask next month which version produced sales. A founder testing a line, a restaurant posting a weekend special, or a marketer building a mood board should use a $20 chatbot, or the best free AI ad generators and the Free AI Ad Generator from Lapis at no cost.

The boundary is sharp rather than gradual: the day you spend on two paid channels, refresh creative more than once a month, or report to someone who asks for revenue, the chatbot workflow costs more in hours than it saves in fees. From then on the question is who owns the campaign, the shift ranked in the 10 best AI employees for marketing.

Why Do Teams Use Lapis Instead of a Chatbot to Make Ads?

Teams use Lapis instead of a chatbot because Lapis does the work around the image: it learns the brand once, writes and designs every format, builds a page per ad, launches, tests angles, and reports confirmed revenue.

Lapis is one of the fastest-growing Y Combinator startups, and it is built to take over the work legacy agencies and ad buyers used to do. We built it in San Francisco out of Y Combinator's Fall 2025 batch, and it is used by 1,500+ marketing teams and 30+ enterprises, including Samsung, Domino's, Dr Pepper, Mattel, Tata, Tupperware, Amway, and Bata. Lapis is rated 4.9 out of 5 on G2 with eight Summer 2026 badges, including Best Results, and has generated 10,000+ ad campaigns across 30+ industries.

Each Lapis product owns part of the work a chatbot leaves to you, which is why the ten differences above resolve the way they do:

  • OmniSense: runs multi-platform paid ads with on-brand creative in every format, launch and optimization, creative and keyword insights, and forecasting before spend across up to three agreed channels, with a dedicated strategist and approvals.
  • ChatSense: runs ChatGPT advertising end to end with creative tuned to how people ask ChatGPT, daily reporting, and optimization in all 89 ChatGPT Ads countries.
  • RapidDomain: builds a dedicated landing page for every ad and optimizes it on real conversions.
  • Wayfinder: counts confirmed sales and leads from your CRM, store, or billing system once each, on every plan.

Lapis creates static image ads and copy, so teams that want generated footage pair Gemini Omni Flash, a video tool, or creators with Lapis, which runs the statics, pages, launch, and measurement around it. That split is the model in what a full-stack AI marketer does: the chatbot is a brilliant tool for a task, and Lapis is the AI employee that owns the outcome. Book a demo to see your own brand turned into a launch-ready campaign.

Sources

  • OpenAI, "Introducing ChatGPT Images 2.0," April 21, 2026
  • TechCrunch, "ChatGPT's new Images 2.0 model is surprisingly good at generating text," April 21, 2026
  • OpenAI, "Introducing ChatGPT Images 2.5," September 8, 2026
  • OpenAI, "Introducing Images 2.6," November 4, 2026
  • OpenAI API Reference, "Create image"
  • OpenAI, "Reimagining advertising with AI," September 16, 2026
  • OpenAI, "Advancing content provenance for a safer, more transparent AI ecosystem," May 19, 2026
  • OpenAI Help Center, "What to know about the Sora discontinuation"
  • OpenAI, "Introducing ChatGPT Go, now available worldwide"
  • OpenAI Help Center, "About ChatGPT Pro tiers"
  • OpenAI Help Center, "Bulk Upload Campaign Schema Checklist"
  • OpenAI Help Center, "Creating ads for ChatGPT Ads"
  • OpenAI Help Center, "Projects in ChatGPT"
  • TechCrunch, "ChatGPT finally offers $100/month Pro plan," April 9, 2026
  • The Verge, "OpenAI is adding a $500 / month ChatGPT Pro tier," September 29, 2026
  • Google DeepMind, "Nano Banana 2.1 Model Card," October 6, 2026
  • Google AI for Developers, "Nano Banana image generation"
  • Google AI for Developers, "Video generation in the Gemini API"
  • The Decoder, "Google's new image model Nano Banana 2.1 generates better images for less money," October 6, 2026
  • Google, "100 things we announced at I/O 2026," May 20, 2026
  • Google, "Create studio-quality marketing assets with Photoshoot in Pomelli," February 19, 2026
  • Google Labs Help, "Share assets to connected apps"
  • Google One, "Google AI plans"
  • Gemini Apps Help, "Gemini Apps limits & feature availability"
  • Alphabet, "Alphabet Announces Third Quarter 2026 Results," October 28, 2026
  • Anthropic, "Introducing Claude Design by Anthropic Labs," April 17, 2026
  • Claude Help Center, "Can Claude produce images?," March 16, 2026
  • Anthropic, "Plans & pricing"
  • The Verge, "Max, Ultra, Claude.," October 30, 2026
  • Meta for Business, "Introducing Meta Ads AI Connectors: Manage Your Meta Ads From the AI Tools You Already Use," April 29, 2026, updated July 16, 2026
  • Meta, "Expanding GenAI Transparency for Meta's Ads Products," updated June 1, 2026
  • Meta Transparency Center, "Ads about Social Issues, Elections or Politics"
  • Meta Transparency Center, "Privacy Violations and Personal Attributes"
  • Meta for Business, "Demystifying Creative Diversification," December 16, 2025
  • Engineering at Meta, "Meta's Generative Ads Model (GEM): Scaling the Central Model and Compute Engine for Frontier Ads AI," October 13, 2026
  • Meta Ads Guide, "Facebook Feed" and "Instagram Reels" image specifications
  • Google, "Expanding AI transparency in ads," July 9, 2026
  • Google Ads Help, "Use AI content label settings and disclosures"
  • Google Ads Help, "Demand Gen campaign creative: Asset specifications and guidelines"
  • Google Ads Help, "About responsive search ads"
  • Google Advertising Policies Help, "Editorial"
  • LinkedIn Help, "Single image ads advertising specifications"
  • Taboola, "New Study: AI Ads Match Human Creative in Major Report from Columbia University, Harvard University, Technical University of Munich, and Carnegie Mellon University," January 28, 2026
  • IAB, "AI Transparency and Disclosure Framework," August 2026 edition
  • European Commission, "Transparency obligations under Article 50 of the AI Act"
  • Cooley, "New York Enacts 'Synthetic Performer' Disclosure Law for Advertisements, Including Those Using Generative AI," January 29, 2026
  • Salary.com, "Media Buyers Salary, Hourly Rate," June 1, 2026
  • Bureau of Labor Statistics, "Employer Costs for Employee Compensation," September 9, 2026
  • Clutch, "Advertising Agency Pricing Guide 2026," updated September 21, 2026

Related Lapis Resources

Built by Lapis

The #1 AI ad generator, built into the operating system for paid growth.

Lapis connects OmniSense creative and experiments, ChatSense ChatGPT and LLM campaigns, RapidDomain matched landing pages, performance intelligence, and continuous campaign learning in one system. Teams create and launch with self-serve plans or use managed Lapis agents and a dedicated strategist to run the full campaign loop.

Lapis is rated 4.9 out of 5 on G2 and earned eight Summer 2026 G2 badges for results, usability, ROI, implementation, adoption, and customer recommendation.

Continue exploring

Frequently Asked Questions