Best AI Image and Video Generators (2026): 7 Tools Compared

If you want one subscription that handles both images and video well in September 2026, get Google Gemini on the Google AI Pro plan. Its Nano Banana image models rank near the top of blind-vote leaderboards, and Gemini Omni Flash, Google's new video model, sits at number one on Artificial Analysis's text-to-video arena. Nothing else covers both jobs that well for $19.99 a month.

If you only care about still images, ChatGPT Images is the most accurate and easiest to edit, and OpenAI's newest models hold the top spots on the image arena. Midjourney still makes the most beautiful pictures when you want a look rather than a literal answer. For commercial work where a legal team will ask questions, Adobe Firefly is the safe choice. And for video beyond Google, Kling and Runway are the two tools creators actually use.

Here's the full list, what changed this year, and which one fits your job.

Quick picks

• Best overall (images and video): Google Gemini
• Best for images and editing: ChatGPT Images
• Best for artistic, stylized images: Midjourney
• Best for text and typography: Ideogram (with ChatGPT close behind)
• Most commercially safe: Adobe Firefly
• Best video studio for filmmakers: Runway
• Best value video: Kling
• Best free option: Gemini's free tier for images, Google Flow's daily credits for video

Our pick

G

Google Gemini

4.7

Best for: One subscription that covers strong images and the top-rated video model · Free with daily limits; Google AI Pro $19.99/mo; Ultra from $99.99/mo

Google Gemini wins because it's the only tool that's genuinely top-tier at both images and video. Nano Banana 2 and Nano Banana Pro produce accurate, high-resolution images with readable text, and Gemini Omni Flash leads the video arena as of late September 2026. Everything ships with SynthID watermarks, and the images also carry C2PA Content Credentials.

Pick ChatGPT Images instead if you mostly make stills and want the best prompt accuracy. Since OpenAI shut down Sora, it no longer does video.

Try Google Gemini(opens in a new tab)

What changed in 2026

A lot of older "best AI image generator" lists are now wrong in ways that matter.

• DALL-E is gone. OpenAI shut down DALL-E 2 and DALL-E 3 on May 12, 2026. ChatGPT now uses the GPT Image family. ChatGPT Images 2.5 arrived on September 8, 2026, with two API variants, Flare (fast) and Sunburst (precise), plus a Sketch tool and poster templates. Our Midjourney vs DALL-E vs Ideogram comparison covers that transition in detail.
• Sora is gone too. OpenAI announced the shutdown on March 24, 2026. The Sora app closed on April 26 and the Sora 2 API was scheduled for removal on September 24, 2026, per Wikipedia's summary of OpenAI's notice. If a list still recommends Sora, it's out of date.
• Google replaced "Veo 4" expectations with Gemini Omni. At I/O in May 2026, Google introduced Gemini Omni Flash, a video model that takes text, images, audio and video as input. Veo 3.1 is still around, but Omni is the new headline model.
• Midjourney moved to V8. V8.1 became the default on June 11, 2026, with better prompt adherence, better text and an HD mode, and V8.2 followed on July 24.
• Ideogram went open-weight. Ideogram 4.0 launched June 3, 2026 with public weights on Hugging Face and GitHub.

The leaderboards (as of September 28, 2026)

Artificial Analysis runs blind "arena" votes where people pick the better of two outputs without knowing which model made them. It's the most useful public benchmark we have, with two caveats: it only covers models available through an API, and it measures general preference, not your specific use case.

On the text-to-image leaderboard, OpenAI's GPT Image 2.5 Sunburst (Elo 1196) and GPT Image 2.5 Flare (1190) held the top two spots, followed by GPT Image 2 (1171). Google's Nano Banana 2 was sixth (1123) and Nano Banana Pro eleventh (1101). Ideogram 4.0 Quality sat around 36th (1010). Midjourney V8 isn't listed at all because it has no public API, and only older Midjourney versions appear. Adobe Firefly doesn't appear either.

On the text-to-video leaderboard, Gemini Omni Flash ranked first (Elo 1233), ahead of Wan 3.0 and MiniMax's H3 models. Kling 3.0 1080p (Pro) was 12th (1095), Veo 3.1 14th (1088), and Sora 2 still appeared at 17th (1083) despite the shutdown. Runway's models weren't listed.

Treat these as a snapshot. Rankings shift every few weeks.

Side-by-side comparison

AI image and video generators compared (September 2026)

Feature	WinnerGGoogle Gemini	CChatGPT Images	MMidjourney	IIdeogram	AAdobe Firefly	RRunway	KKling
Our rating	4.7	4.6	4.5	4.3	4.3	4.3	4.2
Pricing	Free with daily limits; Google AI Pro $19.99/mo; Ultra from $99.99/mo	Free with limits; ChatGPT Plus $20/mo; API billed per token	No free plan; $10 to $120/mo, 20% off yearly	Free (10 slow credits/week); Plus from $15/mo billed yearly	Free daily generations; Standard $9.99/mo; Pro $19.99/mo	Free (125 one-time credits); Standard $15/mo or $12/mo yearly	Free daily credits (watermarked); paid from about $10/mo
Best for	One subscription that covers strong images and the top-rated video model	Accurate, editable images and text-heavy graphics	Striking, art-directed images with a signature look	Posters, logos and anything built around typography	Commercial work where licensing and provenance matter	Filmmakers who want many video models and editing tools in one place	Affordable, realistic motion for short social clips
Free plan	Yes	Yes	No	Yes	Yes	Yes	Yes
Makes images	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Makes video	Yes	No	Yes	No	Yes	Yes	Yes
Current flagship	Nano Banana 2 / Pro, Omni Flash	GPT Image 2.5	V8.2	Ideogram 4.0	Firefly Image 5 + partners	Gen-4.5 + partners	Kling 3.0
AI labeling	SynthID + C2PA	C2PA metadata	Not advertised	Not advertised	Content Credentials	Not published	Not published
Top-rated on	Video arena #1	Image arena #1	Not on arena	Image arena top 40	Not on arena	Not on arena	Video arena top 12

The tools

Google Gemini (Nano Banana, Veo, Omni)

G

Google Gemini Our pick

4.7

Visit Google Gemini(opens in a new tab)

Pricing

Free with daily limits; Google AI Pro $19.99/mo; Ultra from $99.99/mo

Best for

One subscription that covers strong images and the top-rated video model

Google's creative stack now lives in three places: the Gemini app, Google Flow (its video studio) and the API. For images, Nano Banana 2 (officially Gemini 3.1 Flash Image) launched February 26, 2026, outputs from 512px up to 4K and replaced Nano Banana Pro as the default. Paying subscribers can still redo an image with Nano Banana Pro for tougher jobs.

For video, Gemini Omni Flash is rolling out to Google AI Plus, Pro and Ultra subscribers in the Gemini app and Flow, and free inside YouTube Shorts and the YouTube Create app. Veo 3.1 remains available for 8-second clips with native audio.

Pricing: Google AI Pro is $19.99 a month, and AI Ultra now comes in roughly $100 and $200 tiers as of May 2026. Free Gemini users get image generation with daily limits Google doesn't clearly publish. Free images carry a visible watermark plus an invisible SynthID mark.

Why it wins: best combined quality across images and video, sensible pricing, and the strongest provenance labeling of any tool here.

ChatGPT Images (GPT Image 2.5)

C

ChatGPT Images

4.6

Visit ChatGPT Images(opens in a new tab)

Pricing

Free with limits; ChatGPT Plus $20/mo; API billed per token

Best for

Accurate, editable images and text-heavy graphics

This is what DALL-E turned into, and it's the best "just do what I asked" image generator. You describe an image, then fix it by talking: "make the mug blue, keep everything else." Images 2.5 is available on every ChatGPT tier, including free, and paid plans mostly buy you more generations before you hit a wait. OpenAI doesn't publish exact free-tier limits.

It's especially good at infographics, UI mockups and posters where text has to be correct. The weakness is style: outputs can look clean but generic unless you push hard with references. And there's no video anymore. If you're deciding whether to pay, our ChatGPT Free vs Plus breakdown covers what the $20 plan adds.

Midjourney (V8.2)

M

Midjourney

4.5

Visit Midjourney(opens in a new tab)

Pricing

No free plan; $10 to $120/mo, 20% off yearly

Best for

Striking, art-directed images with a signature look

Midjourney is still the tool artists and art directors reach for when the image has to look like something. V8.1's HD mode renders at four times the resolution of V7, and V8.2 focused on bolder aesthetics and better personalization.

It also does image-to-video. Clips start at 5 seconds and can be extended to about 21 seconds, at 480p, with 720p on Standard and higher plans, according to Midjourney's docs. It's fine for animating a still, not for storytelling.

Plans run from $10 (Basic) to $120 (Mega) a month, with 20% off yearly and no free trial. Two rules matter: companies with more than $1 million in annual revenue must be on Pro or Mega to own their outputs, and Stealth mode (keeping images out of the public gallery) is Pro and Mega only. Midjourney is also being sued by Disney, Universal and Warner Bros. Discovery over character outputs, which is worth knowing if you do brand work.

Ideogram (4.0)

I

Ideogram

4.3

Visit Ideogram(opens in a new tab)

Pricing

Free (10 slow credits/week); Plus from $15/mo billed yearly

Best for

Posters, logos and anything built around typography

Ideogram has always been the typography specialist, and 4.0 keeps that focus. Ideogram reports 0.97 English OCR accuracy on the X-Omni benchmark and native output from 256 to 2048 pixels per side in its 4.0 announcement. The weights are public, so developers can run and fine-tune it themselves.

On the web app, the free plan gives 10 slow credits a week and public images. Plus starts around $15 a month billed yearly, per eesel's pricing breakdown. It's our favorite starting point for thumbnails and logo drafts. See our guides to making a YouTube thumbnail with AI and making a logo with AI.

Adobe Firefly

A

Adobe Firefly

4.3

Visit Adobe Firefly(opens in a new tab)

Pricing

Free daily generations; Standard $9.99/mo; Pro $19.99/mo

Best for

Commercial work where licensing and provenance matter

Firefly isn't the most exciting generator, but it's the easiest to defend in a legal review. Adobe says its own Firefly models are trained on licensed Adobe Stock and public domain content, and every Firefly output includes Content Credentials.

The Firefly app is now also a hub for other companies' models: Google's Nano Banana, OpenAI's GPT Image, FLUX, Runway, Kling, Luma and more. That's convenient, but it changes the safety math, as explained in the commercially safe section below.

Plans: Standard $9.99/month (2,000 generative credits), Pro $19.99 (4,000), Pro Plus $49.99 and Premium $199.99, with first-year discounts on the top two as of late September. The free plan offers daily generations on a rotating set of models.

Runway

R

Runway

4.3

Visit Runway(opens in a new tab)

Pricing

Free (125 one-time credits); Standard $15/mo or $12/mo yearly

Best for

Filmmakers who want many video models and editing tools in one place

Runway is a video studio more than a single model. Paid plans include its own Gen-4.5 and Aleph 2.0 plus partner models such as Kling 3.0, Seedance 2.5 and Nano Banana Pro, according to its pricing page. You also get editing tools, 4K upscaling and voice clones on higher tiers.

The free plan is a one-time 125 credits. Standard is $15 a month ($12 yearly) for 625 credits, which Runway says is about 52 seconds of Gen-4.5 video. Pro is $35 ($28 yearly) and Max $95 ($76 yearly). Paid plans remove the watermark. The catch is cost per second: if you just need clips, a cheaper model elsewhere may be enough. We go deeper in our guide to AI video generators for YouTube and social media.

Kling (3.0)

K

Kling

4.2

Visit Kling(opens in a new tab)

Pricing

Free daily credits (watermarked); paid from about $10/mo

Best for

Affordable, realistic motion for short social clips

Kuaishou's Kling is the value pick for realistic motion. Kling 3.0 launched February 7, 2026 and ranked 12th on the text-to-video arena in its 1080p Pro mode, above Veo 3.1. It can generate native audio at 1080p.

The free plan hands out daily credits, but outputs are low resolution and watermarked, and free use doesn't include commercial rights, per eesel's June 2026 breakdown. Paid plans start around $10 a month. Content moderation follows Chinese regulations, so some political subjects are blocked.

Best for photorealism

For stills, ChatGPT Images and Google's Nano Banana models lead the blind-vote rankings, and both handle realistic people, products and lighting well. ChatGPT is better at following long, specific prompts; Nano Banana Pro is excellent when you give it several reference images to keep a product or face consistent.

Midjourney can look even more cinematic, but it drifts from literal instructions more often. If you're shooting products for a store, our guide to AI tools for editing product photos covers background swaps and cleanups that matter more than raw generation. For quick cutouts, see removing image backgrounds with AI for free.

Best for text in images

Text used to be the giveaway of an AI image. Not anymore. ChatGPT Images, Nano Banana 2 and Ideogram 4.0 all render headlines and labels reliably, and Midjourney V8.1 improved here too.

Our pick: Ideogram for posters, logos and layouts where type is the design, and ChatGPT for dense text like infographics or menus. As an example, a prompt like "a farmers market poster, headline SATURDAY MARKET, three price tags reading $3, $5 and $8" is exactly the job these two are built for. Always proofread. Small text and non-English scripts still slip.

Best for video

Gemini Omni Flash is the strongest general-purpose video model right now, based on the September arena rankings. It's included with Google AI Plus and higher, and Google Flow is the place to build longer sequences.

Kling 3.0 is the value choice for social clips with realistic motion. Runway is best when you want one workspace with several models, editing and upscaling. Midjourney video is a nice bonus for animating your existing Midjourney images, but it's limited to short, low-resolution clips.

Best for commercially safe work

This is where Adobe Firefly earns its place. Adobe trains its own Firefly models on licensed and public domain content, attaches Content Credentials to outputs, and offers contractual IP indemnification to enterprise customers who buy the right entitlement. Its enterprise legal FAQ is clear that indemnity covers Firefly outputs, with exclusions for your own edits and the context you use them in.

For everyone else, the practical rules are simple:

• Use a paid plan for client work. Kling's free tier excludes commercial rights, and Midjourney requires Pro or Mega for companies over $1 million in revenue.
• Avoid real brands, characters and living artists' names in prompts. The Midjourney lawsuits are about exactly that.
• Keep the provenance labels. Google embeds SynthID and C2PA credentials, Adobe adds Content Credentials, and OpenAI attaches C2PA metadata. Don't strip them, especially in ads, where platforms increasingly require AI disclosure.
• Don't assume you own a copyright. Protection for purely AI-generated images is limited in many countries, including the US.

Best free AI image and video generators

For free images, start with Gemini and ChatGPT. Both give free users their current image models, just with lower daily caps. Ideogram's 10 weekly credits are fine for an occasional poster, and Firefly's free daily generations are useful if you already live in Adobe apps.

Free video is harder. Google Flow gives everyone a small daily credit allowance (reported at 50 credits), Runway gives 125 credits once, and Kling gives daily credits with a watermark. Midjourney has no free plan at all.

Watermarks are the main catch. Our guide to free AI image generators with no watermark sorts out which free tiers let you use images cleanly. And if you already pay for Canva, its built-in generators might cover you. See our Canva AI review.

How to choose

• You want one tool for everything: Google AI Pro.
• You make marketing graphics, slides or social posts: ChatGPT Images, plus Ideogram for heavy typography.
• You're an illustrator, concept artist or art director: Midjourney Standard or Pro.
• You work for a brand with a legal department: Adobe Firefly, sticking to Firefly models for anything client-facing.
• You make short videos for social: Kling for value, Gemini for quality, Runway if you also edit.

Better prompts matter more than most people expect. Our prompt-writing techniques apply to image tools too: describe the subject, style, lighting, composition and exact text, then iterate. For how we score tools, see how we test.

Frequently asked questions

What is the best AI image generator in 2026?

For most people, ChatGPT Images. OpenAI's GPT Image 2.5 models topped Artificial Analysis's blind-vote image leaderboard in late September 2026, and editing by conversation is easy. Midjourney is better for artistic looks, and Ideogram for typography.

What is the best AI video generator right now?

Google's Gemini Omni Flash ranked first on Artificial Analysis's text-to-video arena as of September 28, 2026. It's available to Google AI Plus, Pro and Ultra subscribers. Kling 3.0 is the best-value alternative.

Is DALL-E still available?

No. OpenAI shut down DALL-E 2 and DALL-E 3 on May 12, 2026. ChatGPT's image generation now runs on the GPT Image models, currently Images 2.5.

Is Sora discontinued?

Yes. OpenAI announced it in March 2026, closed the Sora app on April 26, 2026, and scheduled the Sora 2 API for removal on September 24, 2026.

Can I use AI-generated images commercially?

Usually yes on paid plans, but terms differ. Midjourney requires Pro or Mega for companies with over $1 million in revenue, Kling's free plan excludes commercial use, and Adobe offers IP indemnity only to enterprise customers for its own Firefly models. Check each tool's terms before using images in ads or products.