The standard Glif tests

We run every model through the same prompts. Open one of these 44 tests to see how each model handled it, side by side, or check the leaderboard for how they rank overall.

  • iPhone handheld shot

    26 models

    Handheld phone-selfie camera motion. Looks realistic; the character stays consistent.

  • Miniature diorama

    26 models

    Keeps the tilt-shift miniature look while adding coherent motion. Motions are physically accurate.

  • Animate portrait

    25 models

    Animates the person while preserving identity. Ends on a questioning expression; motion is coherent.

  • Face likeness

    25 models

    The person shown at the end is recognizably the same as in the input image. Temporally stable through the bus pass.

  • Animate radar

    23 models

    Animates the radar module in a realistic way.

  • Action scene

    20 models

    Smooth camera follows the ollie trick. Motion and physics look physcially accurate.

  • Comic book history style

    20 models

    Produces a 4-panel comic layout telling a history-of-the-world narrative. Text and layout are legible. Style and story are consistent throughout the panels.

  • vjou9s0wd7puedqoyxrw.jpgdmba0eqgngycmdc7mgwq.jpgdxxv8hqkoaoqshilis3c.jpgjtbrgfjvi75f6jgnhni4.jpg

    Gibberish prompt

    20 models

    Inspect how a model interprets gibberish prompts.

  • jrormek09xjjcwbklqjp.jpgevujtc4xdy4tchj0jbhj.jpgp10t4xdntukylc3dncnw.jpggi35wmfcgskn4kfxds6z.jpg

    Long and detailed prompt

    20 models

    Adheres to a long and detailed prompt.

  • icfoqimnu4pbh7yqp6g1.jpgmstsovogs4rqqma001bd.jpgrj37ew1ziid5ybjrf7bd.jpgyvbp93xh6clorhrsnyrt.jpg

    Magic laptop

    20 models

    Creatively merges the ancient/mystical elements with a laptop. Runes are present and the screen shows a correctly themed UI.

  • rwedeqpojtt6ymhqlw7u.jpgcdr56ackc0ssebbqzzcl.jpgqvonb6nhmkgzkh3c8wze.jpggqzbndhyyglpm3fa4wnv.jpg

    Miniature scene

    20 models

    Can do tilt-shift scenes. Keeps proportions correct. Correctly produces physically-sound trebuchet. Creates plausible medieval setting.

  • fe3wkmauntdepzby201x.jpgkyhxc85h7ys64ox3ogtr.jpgd69a3w6go9rl5t8335i8.jpgwebjqqncvgnhyl0u8hp2.jpg

    Portrait realistic

    20 models

    Human is realistic and has no artifacts. Applies Kodachrome film effect.

  • rabilbgnsppmijm5wdu5.jpgklqpjgw6dddput8t6msk.jpgs8dryzcy3kvlisvrodew.jpgvkkxv8fq4njobygmoczh.jpg

    Social media selfie

    20 models

    Has correct camera positioning. Looks realistic.

  • lxg3zmj4okzu9qjtatsg.jpgjasv2vnvx5yt5svh6s47.jpgfqpbtt3vhmk5i3k32l8t.jpgv8cu1hsqgogx8jrmt6sa.jpg

    Stacked animals

    20 models

    Stacks the animals in the correct order on top of each other. Keeps the 1980s photo look and the German medieval town setting.

  • qeuualnuszw0pipo4j2d.jpgjztvyp6rte3mlyjmsywh.jpgzcoogmn4zr71jag5swnv.jpggosrqpws9jmjdaq7wxxj.jpg

    Tall forehead

    20 models

    Renders the exaggerated, very tall forehead. No double eyes. Adheres to 1920s black-and-white rubber hose cartoon style.

  • lx8s30x1z0lurpr1q6ql.jpga0typuhfxiw6g6n2ifs8.jpgyjqyzrifughc76vlurr6.jpgiezarjfvp3unxh5to1b3.jpg

    Trading card

    20 models

    Produces a plausible trading-card layout with legible text. Anime style.

  • jigkyiqdqrgaednso0gq.jpgyeujy2sccrv4k2prx75w.jpgmvbjou43bucn9jplcgta.jpguca3jbovyqfr9hspsuuu.jpg

    Text rendering

    20 models

    Adheres to vintage style. Correct lettering and good typography. All info is legible and present.

  • mbqhegkpg5a2ueolaa8w.jpgduj4ugc2ncy3v27vsfwj.jpgzoyqp58upxb7vgg9w7fv.jpgzsdgyecrq2t7meuxlsfb.jpg

    Vague prompt

    20 models

    Inspect how a model handles a vague prompt.

  • xiwzyriwkmzqanmwdyhn.jpgppdgohnezwhotdlyuhjs.jpgb8isuqoz0qu3aiznvavl.jpgnw7fifdkj4oznka73qff.jpg

    CMYK texture

    17 models

    Designs an interesting trading card and applies realistic CMYK dot pattern texture. Applies the texture with the right scaling.

  • xut7hiqgwzqh8hz9er8r.jpgwxfstgu1c4iwnjitwye8.jpgada5kdbxdowk9rhv9ycv.jpgtyktdj960v3vxoewnmrx.jpg

    Cat in space

    16 models

    Adheres to all elements in the prompt. Physically-sound space scene.

  • h7p4oamhyu1xwsd8hssq.jpgwio8mggbrh5wc9qkr9sd.jpgsdl8jvctbqth0oup5thl.jpgll7frxmwhhdl5wpnydvx.jpg

    Character sheet reference

    16 models

    Places the exact same character and using the same style in a scene.

  • rmqrsmesrphmm0nyaqln.jpgydp51ztmhpc8ntljemtt.jpgeeqgwb43sxlswcfrvnnv.jpgbogkgnt7qwkkn9t1c3xt.jpg

    Edit from markup (garden)

    16 models

    Edits only the marked up area. Removes markup. Knows plant names and appearances.

  • ydj8iqhyoaltidd8zbpm.jpgvwfhhohctnykjqkerjjx.jpgwbfy7mmgsa5moeknjyvc.jpgyksxxoabyq5vcwilck2o.jpg

    Edit parts of image

    16 models

    Cybernetics should be perfectly integrated into the subject. Likeness should be preserved. Legible text and annotations.

  • hnpmasf4buq46irrnzaf.jpgen4ba1cfveq5yalfx3ra.jpgbckvhye6n4qsxtjoik53.jpghtktrotjklhqjtc24caq.jpg

    Magazine cover

    16 models

    Uses the exact same character and follows the brief. Good typography and layout.

  • pppr3zlwr97bcn3xqvhz.jpgusc43u61hwsbpobmwwws.jpgsrc0wxsiedhrqc79m1tn.jpgccrpzdcaspjk50ejvp8u.jpg

    Compose image

    14 models

    Combines the referenced woman, white coat and handbag into one coherent photorealistic fashion shot. Identity and products match the references.

  • cnsvf9nz1oe3kh6ktb87.jpgfjotjvectfl2puzu025q.jpgji4galwxjpeptw2jezd8.jpggtcr80bs0z1duptuatrm.jpg

    Style references

    14 models

    Draws a cat cyborg in the style of the reference images, combining features from all the reference images.

  • gag6txytraqttn1a9le8.jpgl3w2ie7rqb0rtiolxytr.jpguo3eetm8zerypnz4yqlh.pngu9omuyvi8dherebqb9nl.jpg

    Product placement

    13 models

    Preserves the reference can exactly (label, colors, proportions). Composes a believable bench scene in Ghent. The can should be sized proportionally.

  • Frame to frame transition watch

    11 models

    Interpolates smoothly from the first to the last frame. The legs are spinning correctly. The watch is spinning in the correct direction.

  • Video to video (anime style)

    5 models

    Turns the man into an anime character while keeping his identity and the original motion. Temporally consistent.

  • ia6r8lemwcwjjz3iuvvv.pngujlhk97ygnmfjtm0hzog.webpfoel7lx33qivx2ssybkt.webp

    Upscale concrete floor

    3 models

    Preserves and sharpens the real-world concrete texture without (VAE) artifacts or invented detail.

  • lnqqtypscfwhawb1o9bc.pngykrkxwyzsfejx7uxv9zx.webpfg39jmkpnx6bu1oql2j4.webp

    Upscale eye

    3 models

    Maintains similar texture.

  • tyzp4k2ghz1cqvyq70lh.pngj4twnmefnxuejvczf4zx.webpdvxhzb9y6vhtmzs5mw99.webp

    Upscale lowpoly

    3 models

    Keeps the low-poly game look; does not hallucinate realistic detail.

  • jvocood8am6nxzxnliz0.pngv1a0avt14xnhxhqmixgf.webpms8xerzmoemicu8aiisd.webp

    Upscale painting

    3 models

    Preserves the watercolor texture; does not turn the painting photographic.

  • rnabcaqdgqnjr9dzfitb.pngquqvwzfa5s7yuto3n2la.webpfiinn71crioljcsgodk3.webp

    Upscale portrait

    3 models

    Adds convincing detail without changing the person's identity. Face and anatomy stay correct.

  • Ambient peaceful

    2 models

    Calm, meditative ambient music with soft piano and strings. Lyrics (if sung) match the lyricsPrompt.

  • Narration calm

    2 models

    Calm, natural narration of the text in the sample voice. No artifacts.

  • Nature rain

    2 models

    Sounds like gentle rain falling on leaves. Realistic; no music or artifacts.

  • ei9v3xydvpd3krhwyuol.png

    Remove background

    2 models

    Clean matte around the hair/curls. Fully transparent background, no leftover fringes.

  • Speech to speech (ElevenLabs)

    2 models

    Speaks the input in the target voice. Natural delivery; content unchanged.

  • Motion control character

    1 models

    Transfers the reference video's motion onto the character while preserving the character's identity.

  • rou8cc2oszwdhsodft0q.jpg

    Outpaint asymmetric

    1 models

    Extends 1024px to the right, revealing more of the room consistently. No visible seam.

  • wvaey7lhrgnd75mo7zid.jpg

    Outpaint uniform zoomout

    1 models

    Zooms out evenly on all four sides. Scene stays coherent; no seams or repetition.

  • wunrx4viesdg7i3bhg0s.jpg

    Outpaint vertical extend

    1 models

    Extends the photo so that more of the man is visible. Extended part should be consistent with the original photo.

  • nir4dqjgeyetkyrrshut.jpg

    Outpaint widen portrait

    1 models

    Extends left and right into a landscape with no visible seams. New content matches lighting and style.