Keling AI Enters the 2.0 Era in Full Force — Kuaishou Finds Its True North at the "Commercial Crossing"

With Keling AI fully entering the 2.0 era, AIGC creators who've been holding their breath can finally exhale. Their only concern now is whether Keling 2.0's "Inspiration Points" will be enough, and which membership tier offers the best value.

As the most closely watched AI creative productivity platform under Kuaishou, the release of Keling AI 2.0 has fundamentally redefined the commercial value and creative boundaries of video and image generation.

Every wild idea you have is redefining the boundaries of AI-powered creation.

Not long ago, we did a complete breakdown of Kuaishou's recent earnings report: Keling AI's cumulative revenue has already exceeded 100 million RMB. Standing on Kuaishou Keling AI's shoulders, we revisited the question: "Kuaishou, who are you? Where did you come from? Where are you going?"

Keling AI has generated over 168 million videos

This time, with the comprehensive upgrade to Keling AI 2.0's foundation models, Kuaishou's ambition becomes even clearer — "An AI commercial stage of infinite possibilities is becoming reality."

Keling AI's key updates cover two foundation models — Keling 2.0 (Master Edition) and Keti 2.0, along with newly launched multimodal editing capabilities.

Keling 2.0 — A Master Emerges, Making Dreams Real

Compared to the Keling 1.6 model, Keling 2.0 (Master Edition) has seen significant improvements in the following areas:

1. Stronger semantic responsiveness — reading between the lines
2. Better dynamic quality — natural, lifelike expressions
3. Superior visual aesthetics — cinematic-grade taste

Source: X user @AiFreak_tool

1) Semantic Responsiveness: Actions, Expressions, Camera Work

Keling 2.0 has achieved notable gains in semantic responsiveness. Compared to Keling 1.6, its natural language processing and prompt comprehension are significantly stronger, enabling it to more precisely capture user intent and generate video and image content that matches expectations.

A large number of AI video creators on X have already started "bombarding" Keling 2.0 with tests and enthusiastically recommending it.

Source: X user @PJaccetturo

The major leap in semantic responsiveness is reflected in Keling 2.0's support for describing more temporally complex shots.

Put simply: it enhances control over video shot time sequences (temporal ordering), enabling the generation of more complex, more dynamic shot effects.

Take the image below, paired with this prompt:

A girl goes from sitting quietly on a park bench to slowly walking out of frame. Morning light gradually shifts to blazing midday sun, then transitions to dusk. Sky colors shift from pink-orange to deep blue to purple-red. Passing pedestrians form flowing ghost trails. Fixed camera, emphasizing light and shadow slowly crawling across the bench's wood grain. Fallen leaves pile beneath the bench, then get swept up by the wind.

Keling 2.0 can generate a smooth video with superior dynamic quality, delivering stunning expressiveness in lighting, color, and dynamic detail.

It's not just simple static frames or single movements — Keling 2.0 can generate multi-segment, continuous, and logically coherent shot effects based on chronological order.

Source: X user @TechByMarkandey

2) Dynamic Quality: Motion Range, Object Coordination, Scene Continuity

Keling 2.0 has also taken a major step up in dynamic generation quality, transforming the video generation experience. Compared to Keling 1.6, it has evolved in motion range, object coordination, and scene continuity, with generated videos approaching cinematic standards.

Keling 2.0 can precisely present actions and camera prompts with smooth, natural transitions, while visual effects and stylistic details appear more refined.

Source: X user @heyDhavall

From a visual presentation standpoint, Keling 2.0 generates videos with strong stylistic consistency and rich, delicate detail rendering, with fairly sophisticated rendering within certain scene layers.

Source: X user @seiiiiiiiiiiru

Keling 2.0 demonstrates high precision in motion capture and camera responsiveness, capable of smoothly processing complex action sequences at highly realistic speeds. The overall dynamic effect comes across as quite true-to-life.

Source: X user @farah_tech_AI

3) Visual Aesthetics: Lighting Effects, Artistic Styles, Image Presentation

Keling 2.0's upgrade in visual aesthetics has pushed the expressive power and beauty of generated content to new heights.

On the visual beauty front, one of Keling 2.0's standout features is its support for 60+ styles, spanning photorealism, cartoon, cyberpunk, watercolor, oil painting, and futurism.

For example, ancient Chinese and modern fantasy urban themes demonstrate Keling 2.0's flexibility and potential across diverse creative subjects.

Source: X user @WuxiaRocks

Upload an image of children running and playing along the beach as waves crash ashore, paired with the prompt:

Oil painting, children running by the sea, waves crashing against the shore.

Keling 2.0 capably captures the children's running dynamics, the rhythm of waves crashing, and the artistic texture of oil painting.

It can understand complex descriptions encompassing artistic styles (e.g., "oil painting"), dynamic elements (e.g., "children running," "waves crashing"), and environmental atmosphere (e.g., "seaside," "sunset").

Keti 2.0 — Multiple Styles, Cinematic Grade

Released alongside Keling 2.0 this time is the Keti 2.0 image model, with several key highlights:

1. Semantic adherence capability
2. Responsiveness to a wide range of styles
3. Cinematic quality

Anyone who regularly uses various text-to-image models has inevitably run into a very practical problem: the longer your prompt gets, with more elements, styles, and detail requirements, the more the model's responsiveness spirals out of control.

This time, Keti 2.0 delivers significant improvements in responding to multiple elements, complex instructions, and color positioning.

When the text prompt is:

A surreal photograph of a river floating out from an oil painting on the living room wall, spilling onto the sofa and wooden floor. The painting depicts a serene river in the mountains. A boat gently rocks in the water, entering the living room. The river's edge spills onto the wooden floor, merging the world of art with reality. The living room is decorated with elegant furniture and a warm atmosphere, cinematic, photograph.

Text prompt:

A banquet hall filled with white tables, surrounded by people enjoying a fine meal.

Text prompt:

Gongbi painting style, Lin Daiyu, wearing Tang dynasty clothing, in the corner of a coffee shop, holding a smartphone in her hand, side backlighting, evening.

Multimodal Editing Capabilities Fully Launched

Beyond the comprehensive upgrades to foundation models, Kuaishou also took a look back at its product offerings. Both Keling 1.6 and the image model now have fully launched multimodal editing capabilities.

Keling 1.6 can now work with video + text/images to replace elements, add elements, and delete elements in videos for better visual results.

Specific capabilities include:

• Replace elements: Specify continuous selection areas across multiple keyframes, using 1 reference image to replace subjects/backgrounds, such as changing clothes or modifying characters.
• Delete elements: Specify selection areas across multiple keyframes to remove subjects/background content, such as removing passersby.
• Add elements: Add subject content through 1-2 reference images or text descriptions, such as adding a spaceship above a city.

The multimodal editing workflow is relatively straightforward.

With the prompt:

Use the panda in [@image] to replace the man in [@video]

Original video:

The person in the video gets replaced with the panda from the image content. Other elements in the scene remain unaffected, with the panda naturally integrated into the video environment while maintaining overall visual coherence and realism.

Modified video:

Launched alongside the video editing feature is also an image editing function — this time, Keling's image generation finally supports inpainting and outpainting:

• Inpainting:
  • Specify the area to repaint, describe target content through prompts to achieve element addition, modification, and other operations.
• Outpainting:
  • Freely adjust outpainting ratios and dimensions, support dragging the original image position and changing size to expand the image content.

If you want the cat in an image to "wear cool sunglasses," simply enter the prompt and specify the repaint area.

Original image, modified area, generated result

The outpainting feature can expand into entirely new areas through simple text prompts, while ensuring natural, smooth results with consistent style.

Simply enter the text prompt:

City street

Original image, outpainting area, generated result

With GPT-4o's Ghibli-style going viral across the internet, Keti 2.0 quickly followed suit, rapidly launching a "style transfer" feature.

For example, 3D cartoon animation style transfer:

Retro beautiful girl style transfer:

Combined with the previous outpainting feature, Keti 2.0 is rapidly iterating and integrating capabilities to build a comprehensive image processing ecosystem.

Beyond the comprehensive upgrades to foundational technology, Keling AI is actively expanding its influence in the AIGC space.

Keling AI hosting the AIGC Film Unit at the 15th Beijing International Film Festival

Keling AI 2.0 "Inspiration Becomes Reality" Creation Competition

Keling AI x League of Legends Mobile AI Initiative

Keling AI Creative Effects Challenge...

And so on, and so forth — all of which demonstrate that Keling AI is not merely a foundation model manufacturer; Kuaishou's product and commercial thinking remains as sharp as ever.

In the AIGC field, simply pursuing a technical route while neglecting commercial pathways means you won't survive the fierce competition for long.

Only by using AI to deconstruct applications, and letting users deconstruct AI — building an ecosystem — can the winds of change disperse the commercialization clouds hanging over AI products.

Different activities target different user groups. Whether professional film and television creators, creative enthusiasts, or gaming content creators — all will become the driving force behind Keling AI's technology-empowered commercial engine.

For AIGC creators, the arrival of Keling AI 2.0 is undoubtedly a feast of inspiration.

From short video bloggers to film production teams, from small e-commerce operations to multinational brands, everyone can find their own commercial opportunity in this wave of technological change.

The only thing left to weigh is perhaps the choice between "Inspiration Points" and membership packages.

After all, when creativity can truly be realized, the only real bottleneck left may be your imagination.