// MODELS
Three tiers. One agent.
Every model in the J@rv1s family runs on the same agentic core. The difference is how much horsepower you point at the problem — and how far you let the agent run on its own.
MJ-1.0
Mini J@r 1.0
fast · lightweight
The everyday agent. Quick reasoning, low latency, optimized for chat, study help, and short automations.
Specs
- Context
- 32k tokens
- Latency
- ~180ms first token
- Modality
- Text
- Best for
- Daily tasks, quick lookups
Features
- Quick reasoning
- Skill toggles
- Project memory
- Low cost per call
JX-1.1
Jarvix 1.1
balanced · multimodal
Mid-tier model with vision, voice, and tool use. Built for research, structured writing, and multi-step tasks.
Specs
- Context
- 128k tokens
- Latency
- ~320ms first token
- Modality
- Text · Vision · Voice
- Best for
- Research, drafting, analysis
Features
- Vision input
- Voice I/O
- Tool calling
- Multi-step planning
JV-1.0
J@rv1s 1.0
flagship · autonomous
Full agentic stack. Browses the web, completes coursework, books appointments, and learns your workflows.
Specs
- Context
- 200k tokens
- Latency
- ~500ms first token
- Modality
- Text · Vision · Voice · Browser
- Best for
- End-to-end automation
Features
- Playwright browser agent
- Long-horizon planning
- Skill composition
- Self-correction
// PICK YOUR TIER
Which one is for you?
Free tier
Mini J@r
Students, daily users
Pro
Jarvix
Researchers, creators
Flagship
J@rv1s
Operators, power users