Kling 4.0 arrives on fal with cinematic control

// 1h agoMODEL RELEASE

Kling 4.0 arrives on fal with cinematic control

Kling 4.0 is now available through fal’s UX in early access, bringing native multimodal video generation, editing, stronger motion stability, and improved character and scene consistency. It supports up to 30-second videos, multiple keyframes, rich reference inputs, stereo audio, and lip sync.

// ANALYSIS

Kling 4.0’s biggest advance is turning video generation from prompt-and-pray experimentation into a more directed production workflow.

–Up to 30-second generation and 10 keyframes enable longer story beats and tighter timeline control
–Up to 15 combined image, video, subject, and voice references improve character and product continuity
–Native stereo audio, dialogue, sound effects, and lip sync make outputs more production-ready
–Video editing and reference-driven replacement reduce the need to regenerate entire scenes
–fal’s UX-first early access is promising for developers, but availability, pricing, and API maturity still need validation

// TAGS

kling-4.0video-genmultimodalvisioninferenceapi

DISCOVERED

1h ago

2026-10-21

PUBLISHED

1h ago

2026-10-21

RELEVANCE

9/ 10

AUTHOR

fal

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE9m ago

Macroscope Adds Scoped PR Approval, Leaner Checks

Macroscope’s Approvability agent now supports onlyApprove path allowlists, limiting automatic approvals to trusted parts of a repository. Check Run Agents also inherit code-review ignore rules, skipping irrelevant files and reducing noise and compute costs.

OPEN SOURCE29m ago

Pele brings dependency-free Mermaid SVGs

Obsidian’s open-source Pele converts Mermaid 12.1 syntax into synchronous, themeable SVG without runtime dependencies. It adds its own layout engine while preserving Mermaid’s familiar diagram language. [GitHub](https://github.com/obsidianmd/pele)

OPEN SOURCE48m ago

TorchTPU Makes Google TPUs Native PyTorch Hardware

Google and Meta open-sourced TorchTPU, a native PyTorch backend for Google TPUs that supports existing workloads with minimal code changes. It brings familiar PyTorch tools, distributed training, torch.compile, vLLM, and Hugging Face integrations to TPU infrastructure.