HUMA-1
H-1
HUMA-1 Default
The flagship model for human-like social behavior, group dynamics, and timing.
Social Reasoning
High
Latency
Real-time (~150ms)
Price
$0.10 / 1k events
Input + Output
Context
Infinite (Vector)
HUMA-1 is our most capable model for handling complex social interactions. It understands group dynamics, knows when to interrupt or stay silent, and maintains long-term relational memory with users. It is designed to be the "behavioral layer" on top of any LLM.
- Natural timing & interruptions engine
- Multi-user group chat awareness
- Long-term memory & context retention
- Event-driven architecture
Pricing
Pricing is based on the number of events processed (input) and actions generated (output). Unlike token-based pricing, we charge per behavioral event to align with the value of social interaction.
Input Events
$0.10
Per 1,000 events
Output Actions
$0.20
Per 1,000 actions
Active Connections
$0.05
Per hour / connection
Capabilities
Social Intelligence
- Turn-takingSupported
- interruptionsSupported
- Sentiment AnalysisNative
- Personality adherenceHigh
Interfaces
- WebSocket (Realtime)Primary
- REST API (Config)Supported
- WebhooksComing Soon
Snapshots
Snapshots allow you to lock your agent to a specific version of the HUMA-1 behavioral engine.
H-1
huma-1
→ huma-1-2025-11-24
huma-1-2025-11-24
Rate limits & Pricing
Rate limits ensure fair usage. Limits are applied per account based on your tier.
| Free | Pay-as-you-go | Enterprise | |
|---|---|---|---|
| Platform Fees | N/A | 5.5% | Bulk discounts available |
| Models | 25+ free models | 300+ models | 300+ models |
| Providers | 4 free providers | 60+ providers | 60+ providers |
| Chat and API Access | |||
| Rate limits | ? | ? | ? |
| Support | Community Support | Email Support | Support SLA with Shared Slack Channel |