Is WEBP smaller than JPG on every image?

That is the general tendency, but it is not a rule without exceptions. The gain varies with the image's content, its dimensions and the chosen quality setting; on some images the difference is pronounced, on others it is quite small. On very small images, WEBP's header information takes up a relatively larger share, so the advantage can disappear and in rare cases even reverse. The only certain way to learn the real gain is to try it with your own images and compare the results.

Is WEBP smaller than JPG in lossless mode too?

That comparison is not meaningful, because JPG has no lossless mode. Lossless WEBP's proper rival is PNG, and on non-photographic content lossless WEBP usually produces a smaller file than PNG. On photographs, lossless mode produces large files in both formats; for photography, lossy mode is the preferred choice.

If WEBP is smaller, why is everyone still using JPG?

The biggest reason is ubiquity and habit. JPG has worked on every device, in every program, everywhere for more than thirty years; it is the default format from cameras to messaging apps. WEBP, although fully supported in modern browsers, can still cause trouble in some older desktop programs and workflow tools. On top of that, converting existing archives takes effort, so JPG stays unchanged in many places.

If I convert the same image from JPG to WEBP, do I get the full gain?

Not entirely, because your source has already been through JPG compression. WEBP's theoretical advantage is at its highest when encoding directly from an original raw image. When you convert a JPG to WEBP, it also has to encode the compression traces JPG left behind, and that reduces efficiency somewhat. Whenever possible, producing WEBP from the original raw file gives the best result.

How Can WEBP Be Smaller Than JPG?

The same picture, fewer bytes: how?

When you save a photograph as both JPG and WEBP, you usually see the same picture on screen, but the WEBP file is typically noticeably smaller. This is not magic; it is the accumulation of compression techniques developed over more than twenty years.

In this article we explain where WEBP gets that advantage from, which design decisions left JPG behind, and what this means in practice.

First, how does JPG work?

JPG was designed in the early 1990s and adopted a relatively simple approach suited to the processing power of that era.

The image is first separated into brightness and color components. Because the human eye is more sensitive to changes in brightness than to changes in color, the color information is usually stored at half resolution — this is called chroma subsampling, and it is where the first loss of information occurs.

Next, the image is divided into fixed blocks of 8x8 pixels. Each block is decomposed into frequency components: slow color transitions inside the block are represented as low frequency, sharp edges as high frequency. During the compression stage, high-frequency components are aggressively rounded off, because the eye notices them least.

This approach has two fundamental limits. First, the block size is fixed — a flat sky and a complex leaf texture both get forced into the same 8x8 mold. Second, each block is encoded on its own, independently of its neighbors.

What does WEBP do differently?

WEBP emerged from adapting techniques developed in video compression to still images. Here are the main innovations it brings:

Block prediction. When WEBP encodes a block, it looks at that block's neighbors and tries to predict its content. If a block in a region of sky closely resembles the block above it, WEBP can say "it resembles the block above, off by this much" instead of encoding the whole block from scratch. Storing just that difference takes far less space than storing the entire block. JPG has no such mechanism; every block is encoded from zero.

Variable block size. WEBP can use blocks of different sizes in different regions of the image. In flat, uniform areas it uses large blocks and covers a wide area with very little data; in detailed, complex regions it drops down to small blocks and preserves precision. JPG's fixed 8x8 structure offers no such flexibility.

More efficient entropy coding. In the final stage of compression, the resulting numbers are packed once more, losslessly. At this stage WEBP uses a more advanced method than the one JPG relies on and squeezes a little more space out of the same data.

The combination of these three techniques means fewer bytes at the same visual quality.

Why does the gain vary from image to image?

WEBP's advantage depends on content, and understanding this matters for setting expectations correctly.

Block prediction works very well on images full of repeating, predictable structures. On photographs containing broad skies, flat studio backdrops or uniform walls, the gain is high.

By contrast, on dense, irregularly textured images where every pixel looks independent of its neighbor — a thick mass of leaves on a tree, say, or a noisy frame shot at high ISO — the prediction mechanism helps less and the gain drops.

On very small images, another factor comes into play: the fixed header information every file carries takes up a relatively large share of the total size and can overshadow the format's advantage.

Transparency and animation: not just size

WEBP's superiority over JPG is not limited to size. WEBP supports the alpha channel, meaning it can store images with transparent backgrounds — something entirely impossible in JPG. WEBP can also operate in both lossy and lossless mode; JPG is lossy only.

This makes WEBP a single format capable of solving the problems that PNG and JPG solve separately. You can use lossless WEBP for a transparent logo and lossy WEBP for a photograph.

WEBP supports animation too, and it is far more efficient than GIF in that area; against GIF's 256-color limit and primitive compression, WEBP offers full color and modern encoding.

What to do in practice

Converting your existing JPG archive to WEBP delivers a meaningful gain, but there is a ceiling: because your source is already lossy, you cannot reach WEBP's full theoretical potential. If you have access to the original raw files, producing WEBP directly from them gives a better result.

Do not pick the quality setting blindly when converting. Try it on a few representative images, examine the result at actual size, and adopt the most aggressive setting where you cannot see a difference. You can use the JPG to WEBP converter to convert your files.

In summary

WEBP's size advantage comes from being able to predict from neighboring blocks, to vary block size according to content, and to use a more efficient final coding stage. This gain is real but it is not a fixed ratio — it varies with content, dimensions and quality setting. Beyond size, support for transparency and a lossless mode also make WEBP a more flexible format than JPG.