Keyframe Labs doubles avatar model resolution, claims 180ms latency at $0.06/min
ycombinator · x · 2026-10-24
Keyframe Labs announced that persona-2.0-live, its real-time interactive avatar model, has doubled resolution (2× resolution, 4× pixels) while keeping word-level control over expression. It's now in preview for all customers via the platform and API.
Key specs from the company:
- Latency: 180ms time-to-first-byte, claimed to beat other providers at 450ms+
- Pricing: as low as $0.06/min, with thousands of concurrent sessions and unlimited call length
- Model family: persona-1.5-live is a 200M-parameter model focused on realism and efficiency; the company frames its work as "human foundation models"
- Capabilities: photorealistic lip sync, natural motion, micro-expressions and director-level emotion control including hair dynamics and lighting physics
More from Models
- 105 planted bugs tested: Kimi K3.1 Pro scores 36, loses on cost-effectiveness — PawelHuryn · 2026-10-24
- Kimi K3.1 isn't an Opus-class model, says investor ahead of detailed benchmarks — bindureddy · 2026-10-24
- Baseten: Mirendil-Base-Large posts lowest perplexity on unseen private code — baseten · 2026-10-24
- Kimi Free Tier Throttled: Users Prompted to Subscribe for Priority Access — ChrisUniverse · 2026-10-24
- Dessa Debuts as a Full-Duplex Speech Model That 'Feels Like Talking to a Person' — Scobleizer · 2026-10-24
- Free tool plots 250+ models' cost curves: prompt caching ROI, GPU breakeven, VRAM sizing — zackku · 2026-10-24