Robot model
MIRA
MIRA is a playable multiplayer world model trained on 10k hours of Rocket League data.
- Manufacturer
- General Intuition
- Form factor
- brain
- Maturity
- research
- Lifecycle
- active
- Website
- generalintuition.com ↗
Overview
MIRA is a playable multiplayer world model trained on 10k hours of Rocket League data. It learns the dynamics of four-player games and runs in real time at 20 fps based on player key inputs. Created in collaboration with Kyutai and Epic Games. General Intuition also builds large action models that generate optimal actions given observations, complementing its world models that predict environment evolution.
Verified vs. claimed
- Maturity stage
- research(Lab-stage only; no commercial or public deployment reported.)
- Sources on file
- View all sources →
No verified deployments on file. DEPLOY tracks independently-sourced deployment evidence separately from manufacturer capability claims.
Key facts
Multiplayer world model (company-stated)
Research paper
Data & sources
Web sources
2
2 sources backing this record.View all →
Pricing
No verified price is on record for MIRA. Physical-AI systems are often sold through enterprise contracts or operated as a service rather than at a public list price.
Recent activity
Every change to this record is dated, sourced, and independently verified where marked.
- Record createdAug 24, 2026
Added to the verified registry
Safety record
No incidents on record for MIRA.
Only active incidents are counted. Retracted incidents are excluded from this summary but remain reachable at their canonical URLs.
Sources (2)
Common questions
What is MIRA?
How much does MIRA cost?
Is MIRA actually deployed in the real world?
Who makes MIRA?
Can you buy MIRA?
What are alternatives to MIRA?
How does MIRA compare to BrainOS?
Is MIRA a top brain?
What is MIRA's maturity stage?
Where is MIRA deployed?
Is MIRA safe?
Methodology: Verified · 2 sources (no primary) · last reviewed 2026-10-10
Verification posture
Verified
Low confidence
Review state
Stable
Last reviewed 2026-10-10
Maturity + lifecycle
Maturity stage: research
Lifecycle: active
Architectural position
Cohort: brain
Sources by quality tier
- 2
- unclassified
- Unclassified source
The framework is documented at /methodology. Corrections at /corrections. Reviewer: DEPLOY editorial team.
Methodology surface for MIRA.Machine-readable surfaces
- Markdown mirror: /models/mira-world-model.md
- JSON-LD: embedded in this page’s head
- REST API: /v1/models/efd5b009-2eb5-4075-b680-7e5279ad15dc
- Revision history: /models/mira-world-model/history
- Data documentation: /data
- Query this programmatically: Deploy MCP
Reality vs attention
Not enough verified signal yet to place MIRA against brain robots. Reality and attention percentiles publish once the underlying record clears its data floor.
6-month trend
Analysis
Thin deployment record; scale validation still in progress.
Signal flags
Dimension breakdown
Verified signal
Attention (reach, not merit)
DEPLOY Intelligence scores are computed from verified registry data: confirmed deployments, disclosed funding rounds, regulatory filings, active job listings, video viewership, and press coverage. Confidence ratings reflect data availability. Scores update nightly.
DEPLOY Indices — verified vs claimed
Last computed: Nov 20, 2026