ImageOptim Alternative for Mac: Smol vs ImageOptim on the Same 24 Photos

Which ImageOptim alternative fits which job?

For photo, screenshot and web-image work, choose Smol: the job is the smallest file that still looks right in a modern format, and ImageOptim cannot do it because it writes neither AVIF nor WebP. For lossless work and SVG, choose ImageOptim.

• Smallest files at a quality you choose: Smol. On our test photos its AVIF output needed 20 to 22% fewer bytes than ImageOptim’s lossy JPEG, and 71 to 76% fewer than ImageOptim’s lossy PNG, at the same quality score.
• AVIF, WebP or HEIC output, converting between formats, resizing: Smol. ImageOptim does none of these.
• PDFs, video and audio in the same queue as your images: Smol.
• Pixel-identical lossless optimization of PNG and JPEG: ImageOptim. Smol always re-encodes.
• SVG: ImageOptim, or SVGO on its own. Smol does not process SVG.
• A budget of zero: ImageOptim.
• One image you want to inspect at several settings before committing: Squoosh, free in the browser.

ImageOptim vs Smol: what each one did to the same 24 photos

Smol’s AVIF output was smaller than ImageOptim’s lossy JPEG at equal quality. ImageOptim kept every pixel and finished JPEGs faster. Here is the test behind that.

The test images are the 24 photographs of the Kodak Lossless True Color Image Suite, a public set that image-compression research has used for decades. ImageOptim keeps whatever format you give it, so we ran everything twice: once from the original PNGs (15,394,305 bytes) and once from the same pictures saved as high-quality JPEGs (4,082,179 bytes), which is closer to what a camera or a design tool hands you. ImageOptim was version 1.9.3, the current release. Smol was 1.0.35. Both ran on the same M2 Pro.

Quality is SSIMULACRA2, a perceptual metric where higher is closer to the original and its authors describe 70 as high quality. Every output was scored against the original picture, not against the JPEG it was made from.

• Out of the box, ImageOptim barely moves a photograph: 8.9% off the JPEGs and 4.3% off the PNGs. That is what lossless means, and it is the reason many people conclude ImageOptim "did nothing" to their photos.
• One row is worth reading twice. From the same JPEGs, Smol’s AVIF at 65 was 17% smaller than ImageOptim’s lossy JPEG at 80 and scored higher (75.6 against 74.3). Smaller and closer to the original at once.
• Photos stored as PNG are where the gap is widest. ImageOptim can only hand you back a PNG; Smol converting the same files to AVIF saved 92.1%.
• ImageOptim won on JPEG speed by a wide margin: about 0.9 seconds for the folder against about 7.1 for Smol. On PNGs it was the other way round.

• Full results and method
• Raw data, 696 scored files (CSV)

At the same quality, which one makes the smaller file?

A compressor can always make a smaller file by throwing more of the picture away, so the table above cannot settle it alone. This is the fair version: ImageOptim with lossy mode on, both tools run at several settings, and the bytes each needs to reach the same SSIMULACRA2 score worked out image by image.

• Kept in the same format, the two are level. If all you will ever want is a smaller JPEG that is still a JPEG, ImageOptim’s lossy mode gives you an equivalent file for free, faster.
• The difference is the format. AVIF reaches the same quality in fewer bytes than JPEG or PNG can, every current browser displays it, and ImageOptim has never written it.
• Smol’s WebP did not beat ImageOptim’s JPEG on this metric. We print that row because it is what we measured.

Is Smol worth $29 if you already use ImageOptim?

Yes, if your images end up in a browser and you want AVIF or WebP. The same 24 photos came out 20 to 22% smaller than ImageOptim’s lossy JPEGs at equal quality, in a format ImageOptim has never written. No, if you only need lossless PNG and JPEG passes or SVG: ImageOptim does that for free.

The terms: $29 once for one Mac, no subscription, no per-file quota, nothing uploaded. There is no free trial and purchases are not refundable, so read the sections on what Smol cannot do before you buy.

• Buy Smol, $29 once for one Mac
• Convert images to AVIF on a Mac
• Compress images for the web
• Smol for WordPress image optimization

Is ImageOptim still maintained? No release since 29 October 2023

The code is; the app is not. ImageOptim’s last release, 1.9.3, shipped on 29 October 2023, and nothing merged since has reached a build you can download.

That date is not an estimate. ImageOptim distributes updates through a Sparkle feed at imageoptim.com/appcast.xml, and that feed contains exactly one item: version 1.9.3, published 29 October 2023. The changelog at imageoptim.com/changelog.html tops out at the same version, the GitHub Releases tab for the project is empty, and the download page still reads "Requires macOS 11+. Works with Sonoma." Sonoma also shipped in 2023. We re-checked all of this on 9 October 2026.

The repository is alive, which is the part people get wrong in the other direction. Community pull requests still land; the most recent commit on main is dated 13 July 2026. What stopped was releases. Check that before you build a workflow on the app.

• ImageOptim update feed

Is ImageOptim slow on Apple Silicon?

On PNGs the released build was slow in our test; on JPEGs it was fast. A merged fix for Apple Silicon exists, and it is not in any build you can install.

On 13 July 2026, pull request #466 was merged with the title "Fix: Apple Silicon performance — stop pinning workers to efficiency cores." Its description explains that ImageOptim’s shipping default of RunLowPriority combined with an unconditional setpriority() call forces every compression subprocess onto Apple Silicon efficiency cores, "which are 3-4x slower than performance cores." On Intel the same setting only meant slightly less CPU time, because every core ran at the same speed. For Apple Silicon the author expects a "dramatic speed improvement."

Here is what the released build did on our M2 Pro: about 21 seconds to optimize 24 small PNGs losslessly and about 17 seconds with lossy mode on. Set to produce lossy PNGs of the same size and score, Smol took about 10.4 seconds, 1.6 times faster. Thorough PNG optimization is slow work in any tool, and we did not build ImageOptim from source, so we cannot say how much of that time the unreleased fix would win back. On JPEGs the released build is quick.

• Pull request #466

Does ImageOptim support AVIF, WebP or HEIC? What it cannot do

No. ImageOptim 1.9.3 handles JPEG, PNG, GIF and SVG, and everything below follows from that format list. One job done well is a defensible position for a free app. It gets expensive when you have a second job.

• No AVIF and no WebP output. ImageOptim writes neither, and AVIF is where the saving in our test came from.
• No HEIC. The format an iPhone shoots in goes in nowhere and comes out nowhere.
• No format conversion of any kind. ImageOptim will make a PNG smaller. It will not make it a JPEG, which is why a photo saved as PNG stays several times larger than it needs to be.
• No resizing. A 6000-pixel-wide camera export stays 6000 pixels wide, however well it compresses.
• No PDF, no video, no audio. That is three more apps for the rest of a normal working week.
• No upscaling.
• No metadata-only mode. Stripping EXIF happens as a side effect of an optimization pass, not as a standalone operation you can run on a folder.
• It rewrites the files you drop on it, in place, and moves the originals to the Trash. Its own 1.9.3 release notes include a "workaround for Dropbox getting confused by quick file renames," which tells you how it works. With lossy mode on, the file in your folder is the lossy one, and the original is one emptied Trash away from gone.

What Smol does that ImageOptim does not

Smol is a Mac app for making files smaller, images first. You drop a folder on it, pick a format and a quality, and it writes the results next to the originals.

It writes JPG, PNG, WebP, AVIF, HEIC and GIF when compressing, and converts between those plus TIFF, BMP and PDF. It resizes while it compresses, up to 32,768 pixels on a side. It reads camera RAW files. The measured results above came from two settings anyone can reproduce: the default, and AVIF at quality 65.

It also goes past images. The same window compresses PDFs at five presets, video to MP4, MOV, WebM or MKV, and audio, and it upscales images 2×, 3× or 4×. For most people that replaces three or four single-purpose utilities.

A separate Neural mode takes a target in kilobytes, anywhere from 4 to 4096, and encodes to hit it. "Get this hero image under 200 KB" becomes a number you type once instead of a slider you nudge and re-export. The measurements on this page are for standard compression, not Neural mode.

If compressing a file would make it larger, Smol keeps the original. You never end up with a worse file than you started with.

When ImageOptim is still the right answer

Three cases, and they are real ones.

The first is lossless work. If the pixels must not change, because the file is a master, a medical or legal record, or an asset someone else will edit, use ImageOptim. Smol has no lossless mode. Expect modest savings on photographs, 4.3% to 8.9% in our test, and far more on flat graphics such as logos and UI screenshots.

The second is SVG. ImageOptim optimizes it and Smol does not process it at all, so if your assets are mostly icons and illustrations this page is the wrong shopping list.

The third is price and speed on plain JPEG. ImageOptim is free, its source is public under GPL-2.0, and on JPEGs it finished our folder about 8 times faster than Smol. With lossy mode on it produces a JPEG as small as Smol’s JPEG at the same SSIMULACRA2 score, and 1 to 8% smaller by SSIM.

Outside those three, the staleness and the format list decide it. A tool that cannot write AVIF cannot produce the smallest file for a web page in 2026, however good its JPEG and PNG work is.

Other ImageOptim alternatives for Mac, compared

Every price, limit and date below was checked on 25 September 2026 against the tool’s own site or repository. Where something is not published, this says so rather than guessing.

One entry needs context. Squoosh is still the best free thing in this category for judging a single image, and it is not simply abandoned either. Asked directly in issue #1472 on 18 May 2026 whether the project was dead, Jake Archibald answered: "The problem is the folks who built it left Google, and Google refused to hand the project over, or maintain it." He added that he still holds commit and publish access and planned to update the JPEG XL and AVIF codecs. As of this writing nothing has landed.

• Smol · $29 one-time, no caps · images out to JPG, PNG, WebP, AVIF, HEIC, GIF, TIFF, BMP and PDF · PDF compression at five presets plus PDF to PNG or JPG · video to MP4, MOV, WebM or MKV · audio · upscaling at 2×, 3× or 4× · target-size compression from 4 to 4096 KB · no SVG, no lossless mode.
• ImageOptim · Free, open source (GPL-2.0) · PNG, JPEG, GIF, SVG · lossless by default with optional lossy quality settings · unlimited local batch · no AVIF, WebP, HEIC, conversion, resize, PDF or video · last release 29 October 2023.
• Squoosh · Free, in the browser, processes locally (its README: "Squoosh does not send your image to a server") · PNG, JPEG, WebP, AVIF and more, with a side-by-side quality preview nothing else matches · one image at a time: batch has been an open feature request since 2022 (issues #1259, #1405, #1479) and is still open · last commit on its development branch: 19 August 2024.
• TinyPNG / Tinify · Free tier: 20 compressions per session, 5 MB per image, 3 conversions per session · paid Web Pro and Web Ultra, billed yearly, raising the per-image ceiling to 75 MB and 150 MB · AVIF, WebP, PNG and JPEG · your files are uploaded to Tinify’s servers · no PDF, no video.
• Optimage · $15 one-time, free tier 24 images per day · JPEG, PNG, APNG, GIF, SVG, WebP, HEIC, PDF and several video formats · its published version history ends at 3.5.0 on 26 November 2021, although the site now offers 3.5.1, a version that history page does not list.
• Compresto · $49 one-time for one Mac, $69 for two · images (PNG, JPEG, HEIC, WebP, TIFF, GIF), video, GIF and PDF · drop zone and automatic folder monitoring · AVIF is not listed · no upscaling listed.
• Zipic · Free tier: 25 images a day after a 7-day Pro trial · Personal licence $29.99 covering two devices · JPEG, PNG, WebP, HEIC and GIF on the basic tier, with AVIF, JPEG XL, TIFF, ICNS, PDF, SVG and APNG on Pro · macOS 13+ · no video.

• Ten Mac image compressors compared
• Squoosh alternatives
• Smol vs TinyPNG

Still 100% local, including the metadata pass

This is usually why someone chose ImageOptim in the first place, so it is worth being blunt: Smol does not upload your files. No account, no server, no upload queue, no processing quota. Everything runs on your own Mac.

That is a real difference from the web tools in the comparison above. TinyPNG is a service: your image travels to Tinify’s servers and comes back, which is also why the free tier caps you at 5 MB per image and 20 compressions per session. Squoosh is the honorable exception among browser tools, since it processes in the page.

Metadata gets its own operation instead of riding along with compression. Strip EXIF, GPS, IPTC and XMP from a whole folder without re-encoding anything, which is what you want when a client asks for the photos minus the coordinates and you have no interest in touching the pixels. ImageOptim has no equivalent standalone mode.

Batch work, watched folders and driving it from an AI agent

Drag a folder of 500 mixed files, images and PDFs and videos together, and Smol processes the lot with per-file progress and per-file savings. You can cancel mid-run, and you can re-optimize individual files afterwards without re-running the batch.

Watched folders handle the jobs you would otherwise have to remember. Point one at your screenshots directory with a preset attached and the task stops existing. For the rest there is a menu bar tray, a Finder Quick Action, and clipboard compression: copy a large image, paste a small one.

Smol also ships an MCP server, so an AI coding agent can drive all of this directly. Connect Claude Code, Codex or Google Antigravity from the "Smol for AI" panel and the agent gets compression, conversion, upscaling and metadata stripping as callable operations, every one of them running locally. For a build step that has to get every asset in a directory under a size budget, an agent calling those operations is a better interface than any window.

• How AI agents drive Smol

What changes if you switch from ImageOptim to Smol

Less than you would expect. Drag and drop works the same way. Metadata stripping is on by default in both apps. Nothing is uploaded in either.

Three habits change. Smol writes to a copy by default, adding a _compressed suffix instead of replacing your original, so a setting you regret costs you nothing. You get a quality setting per format instead of one global preference. And the format menu is long enough that you have to decide what you want, where ImageOptim decided for you.

Smol needs macOS 13 or later on Apple Silicon and macOS 14 or later on Intel, with separate native builds for each. ImageOptim’s download page asks for macOS 11 or later, so on an old Mac it is the only one of the two that runs.

You do not have to uninstall anything. Plenty of people keep ImageOptim for SVG and lossless passes and use Smol for everything that ships to a browser.

Where Smol is the wrong choice

Named plainly, because a comparison page that never loses is not a comparison page.

• Lossless. Smol always re-encodes. For pixel-identical output use ImageOptim.
• SVG. Smol does not process SVG. ImageOptim does, and SVGO on its own does it too.
• Raw speed on JPEG. ImageOptim was about 8 times faster on our folder of JPEGs.
• A budget of zero. If $29 is the obstacle, ImageOptim is free and good at what it does, and Squoosh is free and good at something else.
• One hero image you need to agonize over. Squoosh’s side-by-side preview is a better tool for that than any batch app.
• Windows or Linux. Smol is macOS only.
• Build-time optimization on a server. If images are processed during a CI run on Linux, run an image library in the build. A Mac app is the wrong layer for that.
• RAW editing. Smol reads CR2, NEF, ARW and DNG, but it compresses and converts them. It does not develop them. Lightroom Classic and Capture One own exposure, color and lens correction.

Smol vs ImageOptim, feature by feature

Frequently asked questions

Is Smol better than ImageOptim?

For making photos and web images as small as possible, yes, and we measured it. On 24 public test photos, Smol’s AVIF output needed 20 to 22% fewer bytes than ImageOptim’s lossy JPEG and 71 to 76% fewer than its lossy PNG at the same SSIMULACRA2 quality score. Smol also converts formats, resizes, and handles PDF, video and audio. ImageOptim is better for lossless optimization, SVG and price: it is free.

Does ImageOptim support AVIF, WebP or HEIC?

No. ImageOptim 1.9.3 reads and writes PNG, JPEG, GIF and SVG, and it does not convert between formats. It cannot write AVIF or WebP and does not open HEIC. Smol writes all three on a Mac, converts between them in batches, and runs locally like ImageOptim does.

Why did ImageOptim barely shrink my photos?

Because its default mode is lossless: it repacks the file without changing a pixel, and a photograph has little slack to remove that way. In our test it made 24 JPEG photos 8.9% smaller and the same photos as PNG 4.3% smaller. Turning on lossy mode in its Preferences saved 63.8% on the JPEGs, and a lossy compressor such as Smol saved 69.3%.

Is ImageOptim still maintained?

The code is; the app is not. ImageOptim’s Sparkle update feed and changelog both end at version 1.9.3, published 29 October 2023, and its GitHub Releases tab is empty. Community pull requests are still being merged, with the most recent commit on main dated 13 July 2026, but none of that work has reached a downloadable build. Checked 9 October 2026.

Is ImageOptim slow on Apple Silicon Macs?

On PNGs it can be; on JPEGs it is not. Merged but unreleased pull request #466 says the shipping app pins its compression work to efficiency cores, which its author calls "3-4x slower than performance cores." On an M2 Pro, ImageOptim 1.9.3 took about 21 seconds for 24 small PNGs losslessly and about 17 in lossy mode, against about 10.4 for Smol at a matching setting. We did not test the unreleased fix.

What is the best ImageOptim alternative for Mac?

It depends on what ImageOptim is not doing for you. For AVIF, WebP and HEIC output, format conversion, resizing, and PDFs, video and audio in one app, Smol at $29 one-time. For judging a single image at several settings, Squoosh, free in the browser. For a free app with a quality slider and WebP output, Caesium. For lossless PNG and JPEG batches at no cost, nothing has replaced ImageOptim itself.

Does Smol handle SVG or lossless compression the way ImageOptim does?

No to both. Smol does not process SVG, and it has no lossless mode: it always re-encodes the image. If your work is mostly icons and illustrations, or the pixels must stay identical, ImageOptim is the right tool. Smol writes copies by default, so you can run both without it touching what ImageOptim has optimized.

Is Smol a subscription, and does it upload anything?

Neither. Smol is $29 once, including future updates, with no renewal, no per-file quota and no daily cap. Every file is processed on your own Mac, with no account, server or upload queue. There is no free trial and purchases are not refundable.