Runware adds Google DeepMind’s Gemini Omni Flash to its API

Runware now gives developers and enterprises immediate API access to Google DeepMind’s Gemini Omni Flash, a multimodal video model that can generate and edit video from text, images, video and audio. The launch removes waitlists and allows commercial use from day one through a single integration.

Why it matters: - Runware is positioning Gemini Omni Flash as a production-ready video model for teams that want one API for generation, editing, and deployment. - The launch could reduce the need for separate video creation and editing pipelines. - Commercial use is permitted from day one, which matters for companies that need to ship paid products quickly.

What happened: - Runware launched developer API access to Google DeepMind’s Gemini Omni Flash on June 30, 2026. - The model is available now with no waitlist. - Developers and enterprise teams can use Gemini Omni Flash through the same Runware integration that covers the company’s other models. - The launch page, documentation, pricing, and enterprise contact details are available through Runware’s site: Launch the model, Documentation, Pricing, and Enterprise.

The details: - Gemini Omni Flash is the first model in Google DeepMind’s Omni family. - Google describes the model as its most advanced multimodal video generation and editing model to date. - The model accepts text, image, and video inputs. - The model produces video grounded in Gemini’s knowledge of the real world, including physics, geography, culture, and language. - The core capability is conversational, multi-turn editing. - Each edit builds on the last. - Characters stay consistent across scenes. - Locations carry over between edits. - Runware says the model supports text-to-video, image-to-video, reference-to-video, multi-turn video editing, native audio sync, location and world awareness, character and object swaps, sketch-to-video, motion and style transfer, and production at scale. - Native audio sync keeps on-screen text and sound aligned with the visual output. - Location and world awareness is designed to produce regionally accurate settings, signage, and language without manual prompting. - Runware’s API also gives access to its full library of image, video, audio, 3D, and LLM models. - Runware runs on its proprietary Sonic Inference Engine. - The company says the platform uses usage-based pricing, high concurrency, no GPU provisioning, and no long-term contracts. - Runware says its infrastructure supports production-scale workflows behind a single API. - For enterprise teams, Runware offers dedicated requests-per-minute allocation, volume-based pricing, 24/7 engineering support, and one API across its full model library. - Runware is SOC 2 and ISO 27001 certified and GDPR compliant. - The company also supports SSO and SAML for enterprise authentication.

Between the lines: - Runware is trying to make advanced video generation feel like a standard developer service rather than a specialized media workflow. - Multi-turn editing is the standout feature because it reduces the need to regenerate full videos after each change. - The company is also using pricing, concurrency, and enterprise controls as part of the pitch, not just model access. - The launch broadens Runware’s appeal to both individual developers and larger teams that need compliance and support.

What's next: - Runware is directing teams to contact sales for capacity and onboarding discussions. - Enterprises interested in production usage can ask about dedicated RPM allocation and volume pricing. - Developers can start integrating the model immediately through Runware’s documentation and API endpoints.

The bottom line: - Runware now offers Gemini Omni Flash as a commercial, API-accessible video model with editing features, enterprise controls, and a single integration alongside the rest of its model catalog.