Synthesia’s 3.0 update, announced in October 2025, is now rolling out key features in January 2026. The biggest immediate change: Express-2 avatars are now unlimited across all plans, removing one of the most frustrating limitations from previous versions.
Express-2 avatars go unlimited
Previously, avatar access was tiered by plan level – pay more, get more avatar options. That’s gone. Every Synthesia plan now includes unlimited access to Express-2 avatars. If you’ve been stuck with a limited avatar roster or paying for higher tiers just to access specific looks, this levels the playing field considerably.
Express-2 avatars are Synthesia’s higher-quality models with better facial expressions, more natural movements, and improved lip-sync accuracy. Making these unlimited means even starter-tier users can create professional-looking video content without avatar limitations holding them back.
Veo 3 and Sora 2 integrations
Synthesia integrated Google’s Veo 3 and OpenAI’s Sora 2 for B-roll footage generation. You can now prompt avatars to perform specific actions – walking, driving, cooking, or any task you need demonstrated. Previously, avatars were mostly static presenters. Now they can actively demonstrate processes.
This matters for training videos and product demonstrations. Instead of finding stock footage of someone using software or performing a task, you can generate an avatar doing exactly what you need. The prompt-based approach means you describe the action and the system generates appropriate footage.
Both integrations are accessible through the new AI Playground feature, which appears as a sidebar tab. The Playground lets you experiment with the latest generative AI models directly within Synthesia’s interface. You’re not just locked into Synthesia’s own models – you can test what Google and OpenAI’s tools can do for video generation and pick what works best for your specific use case.
What’s coming next
Video Agents for Enterprise are expected early 2026. These will be interactive video experiences that can respond to user inputs, creating personalized training or onboarding flows rather than static videos.
Copilot, planned for 2026, promises to function as a professional video editor that can write scripts in seconds, connect to knowledge bases, and suggest visual elements. If it delivers, this could significantly reduce production time for routine corporate videos.
Courses, also on the 2026 roadmap, aims to transform workplace learning with interactive experiences using avatars, video agents, and built-in interactivity. The goal is measuring skill development, not just video completion rates.
Current capabilities at scale
Beyond the 3.0 updates, Synthesia’s existing feature set remains substantial: 240+ avatars, 160+ languages, AI dubbing with lip-sync across 30+ languages, and 1-click translation to 80+ languages. The Quizzes feature helps improve video completion rates by making content interactive.
The company raised $200M in October 2025 at a $4B valuation, signaling investor confidence in the enterprise video generation market. That funding likely accelerates the roadmap features mentioned above.
Who benefits most
Enterprise training teams get the most immediate value. Unlimited Express-2 avatars mean you can create diverse, professional training content without budget constraints on avatar access. The Veo 3 integration for task demonstrations makes procedural training more practical to produce at scale.
Multilingual organizations will appreciate that avatar quality improvements extend across all languages. Creating training content in 30+ languages with consistent quality used to require massive localization budgets. Synthesia’s dubbing with lip-sync makes this economically feasible.
Content teams producing high-volume explainer videos can now experiment with different avatar styles and B-roll approaches through the AI Playground without committing to a single generation method. Test what works, then scale production.
Production workflow impact
The unlimited Express-2 access removes decision friction. You’re no longer calculating whether a specific video justifies using a premium avatar. Pick the avatar that fits the content, generate the video, move on. This sounds minor but adds up significantly when producing dozens or hundreds of videos.
Veo 3 B-roll generation addresses a common pain point: finding or creating footage that matches your script exactly. Stock footage rarely matches perfectly, and custom filming is expensive. Generating an avatar demonstrating the specific task you’re explaining keeps everything aligned.
Worth upgrading?
If you’re already on Synthesia and frustrated by avatar limitations, the unlimited Express-2 access alone justifies staying current with 3.0. The quality difference between Express-2 and earlier avatars is noticeable, and having unrestricted access changes what’s practical to produce.
For teams evaluating Synthesia for the first time, the combination of unlimited avatars, AI Playground experimentation, and upcoming Video Agents makes this a stronger offering than previous versions. The platform is evolving from “generate talking head videos” to “create interactive, personalized video experiences at scale.”
The roadmap features (Copilot, Courses) suggest Synthesia is positioning for the full video production workflow, not just avatar generation. If those deliver as promised, this becomes a comprehensive video platform for enterprise learning and communication.
Current users should update to access unlimited Express-2 avatars immediately. New users should evaluate now if the roadmap features align with their 2026 video strategy – early adoption might provide advantages as Video Agents and Copilot roll out.
