Side by side
Spatius vs Tavus CVI
Any number of models of one kind, normalised onto the same cost basis and scored through the lens you pick.
| How it looks | ||
| Cost of talking | ||
| Per minute | $0.0072 | $1.00Estimate |
| Per hour | $0.432 | $60.00 |
| 30-min interview | $0.216 | $30.00 |
| List price | $0.0072/minPublished | $1/minEstimate |
| Recruitment fit | ||
| Overall | 80/100 | 90/100 |
| Verdict | A very cheap, good-looking avatar that runs in the browser, without a visual perception layer. | The most complete option for conversational screening, the perception layer is worth the premium when a real person is on the other end. |
| Best for | Browser-based interviewsAlways-on job page avatars | First-round conversational screeningEmployer-brand videos personalised per candidateProctored assessments |
| Watch out for |
|
|
| What it animates | ||
| Coverage | ||
| Face | Full | Full |
| Lip-sync | Full | Full |
| Upper torso | Partial | Partial |
| Hands | None | None |
| Full body | None | None |
| Prompt artifacts | None | None |
| Capabilities | ||
| Modalities | avatar | avatar, video, speech |
| Context window | n/a | n/a |
| Variants | Real-time | Phoenix-4.5, Phoenix-4, Phoenix-3 (legacy) |
| On HumanLike | Interview | AI Personas, Interview |
| HumanLike evals | ||
| Wired into HumanLike avatars | n/a | Yes, face_id providerOur eval |
| Published benchmarks | ||
| Lip-sync accuracy | n/a | 90%Estimate |
| Utterance-to-utterance latency | n/a | 1000 msEstimate |
Cost per hour assumes 150 wpm and a 40% AI speaking share. Hover a figure for its full derivation.