Models¶
Moveris offers Mixed V2, Mixed V3, Mixed V3.1, and Mixed V3.1.1 model families to match different integration needs. The comparison table below covers Mixed V2; see Mixed V3 Models, Mixed V3.1 Models, and Mixed V3.1.1 Models for newer generations.
| Fast | Balanced Recommended | Thorough | Extended | Maximum | |
|---|---|---|---|---|---|
| Model ID | mixed-10-v2 | mixed-30-v2 | mixed-60-v2 | mixed-90-v2 | mixed-120-v2 |
| Frames | 10 | 30 | 60 | 90 | 120 |
| Capture time | ~0.3 s | ~1 s | ~2 s | ~3 s | ~4 s |
| Response time | ~1 s | ~3 s | ~5 s | ~7 s | ~10 s |
| EER | 4.4% | 4.0% | 4.7% | 5.7% | 4.6% |
| AUC | 0.988 | 0.991 | 0.991 | 0.989 | 0.991 |
| Balanced accuracy | 95.2% | 95.7% | 95.2% | 94.3% | 94.7% |
| Best for | Low-friction, high-volume flows | Standard KYC & identity verification | High-security onboarding | Escalation, compliance-heavy flows | Highest scrutiny, regulatory edge cases |
Model cards
Detailed model cards with full metrics and evaluation data are available in the Developer Portal. Sign in to access them.
Not sure? Start with Balanced (mixed-30-v2) or V3.1.1 (mixed-30-v3_1_1)
Mixed V2: mixed-30-v2 delivers strong overall accuracy (4.0% EER, 0.991 AUC) with ~1 second of capture.
Mixed V3.1.1: mixed-30-v3_1_1 is the latest 30-frame generation. Prefer V3.1.1 for new integrations when your deployment supports it.
Mixed V3 Models¶
Mixed V3 is the next-generation model family, available at 10, 30, and 60 frames.
| Fast (V3) | Balanced (V3) | Thorough (V3) | |
|---|---|---|---|
| Model ID | mixed-10-v3 | mixed-30-v3 | mixed-60-v3 |
| Frames | 10 | 30 | 60 |
| Notes | Often visual-only | Often visual-only | Full temporal-physio fusion |
V3 models use a temporal-physio architecture: visual analysis combined with physiological signals (e.g. heart-rate cues) across frames. Reliable physiological extraction needs enough temporal coverage, so mixed-60-v3 runs full fusion. At 10 or 30 frames the physiological signal is often unavailable, and the model falls back to a visual-only prediction (flagged via the response warning field). Use V3 when your deployment supports it.
Mixed V3.1 Models¶
Mixed V3.1 is the latest model generation. Four frame counts ship in production: 10, 30, 60, and 90.
| 10f | 30f | 60f | 90f | |
|---|---|---|---|---|
| Model ID | mixed-10-v3_1 | mixed-30-v3_1 | mixed-60-v3_1 | mixed-90-v3_1 |
| Frames | 10 | 30 | 60 | 90 |
| Recommended for | Lowest latency, new integrations | Default V3.1 choice | Best attack rejection (V3.1) | Extended temporal coverage |
V3.1 vs V2
V3.1 models use a different architecture and scoring regime than Mixed V2. Pin an explicit alias (mixed-30-v3_1, not mixed-30-v2) when migrating. See Model Versioning & Frames.
Mixed V3.1.1 Models¶
Mixed V3.1.1 is the latest model generation and a clean retrain of the V3.1 temporal-physio family. Five frame counts ship in production: 10, 30, 60, 90, and 120.
| 10f | 30f | 60f | 90f Recommended | 120f | |
|---|---|---|---|---|---|
| Model ID | mixed-10-v3_1_1 | mixed-30-v3_1_1 | mixed-60-v3_1_1 | mixed-90-v3_1_1 | mixed-120-v3_1_1 |
| Frames | 10 | 30 | 60 | 90 | 120 |
| Recommended for | Lowest latency, new integrations | Default V3.1.1 choice | High-security onboarding | Best overall accuracy (ship config) | Maximum temporal coverage |
V3.1.1 was retrained on clean, canonical Red-Green-Blue (RGB) inputs and recalibrated on the real-world holdout set, where the equal error rate (EER) drops to roughly 5–9%. The architecture matches V3.1, so the same frame-count guidance applies: more frames give the temporal-physio model more signal. Moveris ships mixed-90-v3_1_1 as the recommended configuration.
V3.1.1 vs V3.1
V3.1.1 is the recommended path for new integrations. Pin an explicit alias (mixed-30-v3_1_1) and migrate from mixed-N-v3 or mixed-N-v3_1 when your deployment supports V3.1.1. See Model Versioning & Frames.
When to Use Each¶
Fast — Users are impatient, risk is low, or you're re-verifying someone already trusted.
Balanced — Standard identity verification, account creation, financial services.
Thorough — A single false accept would be costly. Users will tolerate a longer check.
Extended — Compliance-heavy flows, escalation when Thorough is borderline, or when regulators expect longer temporal coverage (~3 seconds).
Maximum — Highest scrutiny scenarios: regulatory edge cases, high-value transactions, or final-level escalation when you need maximum temporal redundancy (~4 seconds).
Escalation Pattern¶
You don't have to pick just one. Use Fast by default and escalate when needed:
User submits → Fast (mixed-10-v2 or mixed-10-v3_1_1)
→ High confidence → Done
→ Borderline → Balanced retry (mixed-30-v2 or mixed-30-v3_1_1)
→ Still borderline → Thorough (mixed-60-v2, mixed-60-v3, or mixed-60-v3_1_1)
→ Escalation → Extended / Maximum (mixed-90-v2, mixed-120-v2, mixed-90-v3_1_1)
How Frames Affect Results¶
More frames give the model more temporal signal to analyze — more micro-expressions, more biological coherence data across time. All v2 models achieve strong accuracy (94–96% balanced), so the primary tradeoff is latency vs. temporal redundancy rather than a large accuracy gap.
All models use the same API endpoints. To select a model, pass the model ID in your request:
Frames minimum (recommended exactly)
The API requires at least the number of frames the model expects. mixed-10-v2 expects 10 frames, mixed-30-v2 expects 30, mixed-60-v2 expects 60, and so on. For predictable latency, send exactly the required number.
Deprecated Models¶
The legacy mixed models (mixed-10, mixed-30, mixed-60, mixed-90, mixed-120, mixed-150, mixed-250) are deprecated. Migrate to the corresponding mixed-N-v2 variants (e.g. mixed-10 → mixed-10-v2).
See Model Versioning & Frames Parameters for alias behavior, version pinning, and the exact frame rules per endpoint.