Jackson Wang

Grok is constantly being updated, so there is a good chance that what didn’t work for you even a few days ago might work now

🚨 xAI quietly made Grok Imagine much easier to use. You can now multi-select images and videos, tag them in groups, and bulk download or bulk delete selected files. A huge quality-of-life upgrade if you generate a lot or do video editing.

Good quality, and knowledge coverage. Try some crazy ideas that you didn't think was possible before, and let us know how we can continue to be better!

happy easter from the @grok imagine team to you :))

Grok Imagine Quality mode can make beautiful infographics in record time 🔥 { "subject": "TERAFAB, a proposed semiconductor complex by Tesla, SpaceX, and xAI", "intent": "beautiful high-end infographic grounded in semiconductor manufacturing, not a generic robot factory", "mode": "concept announcement visual, not a finished documentary photo", "canvas": { "aspect_ratio": "16:9", "resolution": "ultra", "background": "deep black graphite with faint cleanroom haze, subtle film grain, controlled vignette" }, "headline": { "title": "TERAFAB", "subtitle": "SEMICONDUCTOR SYSTEM FOR EARTH + ORBIT", "placement": "center top", "font_style": "clean geometric sans-serif, ultra precise, Tesla-like restraint", "color": "#F3F4F6", "tracking": "wide", "hierarchy": "strong title, restrained subtitle" }, "composition": { "layout": "strict symmetrical center composition with two mirrored output streams", "centerpiece": "one monumental silicon wafer and chip-stack diagram floating in the middle", "left_stream": "terrestrial inference chip path", "right_stream": "space-hardened chip path", "lower_band": "under-one-roof manufacturing stages", "negative_space": "generous and intentional, no clutter" }, "central_visual": { "type": "hybrid semiconductor sculpture + technical infographic", "description": [ "a large 300mm silicon wafer viewed front-on, polished dark mirror surface with fine die-map geometry", "a precise cross-section through the wafer revealing layered chip architecture", "photomask patterns, circuit traces, microscopic interconnect grids, package substrate layers, test contact points", "floating chip package stack beneath the wafer, beautifully aligned, clean, engineered, monumental", "no car assembly line, no factory robots, no generic sci-fi machinery" ], "materials": [ "polished silicon", "brushed aluminum", "clean glass", "dark ceramic package surfaces", "fine copper interconnect accents" ], "lighting": "cold white cleanroom beams from above, sharp edge reflections, restrained glow, no neon excess" }, "dual_chip_story": { "left_path": { "theme": "EARTH", "label": "EDGE INFERENCE", "visual_language": [ "copper-gold signal flow", "subtle silhouettes of a vehicle control board and a humanoid hand actuator", "dense efficient compute, compact power envelope", "clean circuit routes moving from wafer to packaged chip to terrestrial deployment" ] }, "right_path": { "theme": "ORBIT", "label": "SPACE-HARDENED COMPUTE", "visual_language": [ "cold blue-white signal flow", "subtle satellite geometry, orbital arc lines, thermal radiator motifs", "high-temperature and radiation-tolerant implication without cartoon effects", "clean circuit routes moving from wafer to packaged chip to orbital deployment" ] } }, "manufacturing_band": { "placement": "lower third", "style": "elegant linear process strip with exact spacing", "stages": [ "DESIGN", "MASK", "LITHOGRAPHY", "FAB", "MEMORY", "ADVANCED PACKAGING", "TEST" ], "visuals": [ "mask reticle geometry", "cleanroom corridor modules", "wafer carriers", "process chambers as abstract engineered forms", "metrology markers", "package cross-sections" ] }, "location_and_scale": { "placement": "small controlled annotations", "elements": [ "AUSTIN, TEXAS", "TWO CHIP FAMILIES", "TERAWATT-SCALE TARGET", "UNDER ONE ROOF" ], "style": "tiny monochrome microtext, highly legible, evenly spaced" }, "typography": { "system": "single geometric sans-serif family plus minimal monospaced micro labels", "rules": [ "perfect kerning", "high readability", "no warped glyphs", "no gibberish text", "clear hierarchy through size and spacing only" ] }, "graphic_elements": { "connectors": "hairline technical rules, exact anchor points, no floating lines", "grids": "very subtle engineering grid in background", "micro_markers": "measurement ticks, die labels, alignment marks, tiny index numbers", "icons": "minimal semiconductor and orbital symbols only" }, "color_palette": { "base": [ "#07090B", "#111317", "#F3F4F6", "#9CA3AF" ], "accents": [ "#C58B3A", "#7CC7FF" ], "usage": "gold-copper only for terrestrial signal path, cold blue only for orbital signal path, everything else neutral and disciplined" }, "aesthetic_direction": { "style": "luxury industrial editorial, launch-event graphic, advanced semiconductor beauty", "mood": "inevitable, precise, ambitious, engineered", "finish": "crisp, expensive, calm, exact" }, "strict_constraints": [ "do not depict a generic autonomous megafactory", "do not use automotive assembly lines", "do not use random robot arms as the main subject", "do not turn it into a space colony poster", "do not emphasize solar farms or construction equipment", "do not use fake dashboard UI clutter", "do not overload with unreadable statistics", "do not make it look like a stock corporate infographic" ], "forbidden": [ "blurry text", "plastic materials", "rainbow sci-fi lighting", "busy composition", "fantasy machinery", "solar-panel-dominant imagery", "cartoon satellites", "random meaningless numbers" ] }

@xdNiBoR The future of AI is primarily video understanding and generation, because photons are by far the highest bandwidth form of communication. These are essential tools for AGI. Worth mentioning that Imagine is positive gross margin for @xAI, not a money loser.

Thrilled to announce that Canvas-to-Image (C2I) is accepted to #SIGGRAPH2026! 🎥 We simplify complex generation into one intuitive interface. Control ID, Pose, and Boxes: all from a single canvas. You are the designer. Kudos to the team: @guocheng_qian, Maya Goldenberg, @tsaishien_chen, @AbermanKfir, @SergeyTulyakov, @PINguAR and @kcjacksonwang! See you in LA!

Truly proud of what a small team was able to achieve at @xai. Ad astra!

@alilbitofmo @xai 🐐

Today was my last day at @xai . The past year (and some change) have been nothing short of exceptional—inspiring and grueling in equal measure. It was an honor to support the Imagine team from the human data side, exhilarating to champion and deliver impossible outcomes on impossible timelines, and deeply inspiring to see our work improve the model in real time. If I didn't get to thank you personally today, here's my heartfelt thanks to the Omni/Imagine team, the Image & Video Human Data crew, and all my colleagues who slogged through the trenches with me day after day (across eng, HD, and beyond). It was a true honor working alongside you—I’ll miss you all very much. Excited for whatever comes next, and rooting for the xAI team! 🚀 Ad astra, etc.

@nonieengel @kennylugo11 @hexiang chibichu 🥰

@kennylugo11 so cute! @hexiang , @kcjacksonwang, where are your chibis 👀 waiting...

Turned my wife and I into Chibis Made with Grok Imagine Chibi template grok.com/imagine

I love Chibi 💖

Follow @imagine to be the first to know cool updates! 😎

Proud of our team that makes the huge leap happen compared to last version but this is just the start. Better models are lined up and we keep improving every week. Join us towards Superhuman Multimodal Intelligence job-boards.greenhouse.io/xai/jobs/50826… !!

Since @Grok Imagine v1.0 release, many of you requested video extension. We hear you! It's available today 🎬 Enjoy, with more to come!!

seek truth in abstraction, pt. II

The @xai team is at Stanford’s Meyer Green from 1-4pm today!! Come say hi, we’d love to meet Thank you @BASES for hosting the career fair

Despite all the crazy advances in video generation, I still don't see models that can combine elements from "multiple videos". Probably the next big thing .. meanwhile, check out Max's work.

As we move towards AI data being better than average real data, the skill of using frontier model and synthesizing new data to enable better control is gold. Congrats Rishubh!

Use multi-image editing in Grok Imagine to combine styles, characters, environments, and more. Share your creations in the comments.

Imagine stands as one of the main achievement we accomplished in xAI, and we are growing our team! If you agree with our mission, join us: - Image Agent: job-boards.greenhouse.io/xai/jobs/50132… - Image Engine: job-boards.greenhouse.io/xai/jobs/46816…