Primary use stated or inferred cautiously from official documentation.
Invoker-Render 1.0
Causal renderer for Code World Model 2.0 with 33.12 billion parameters. Its weights were published on November 16. It converts structural proxies, an appearance reference and text into video, while another engine maintains world state.
The essential
Maximum input capacity when published by the source.
Documented maximum generation limit.
Verified availability channels.
Primary functional family and verified specialties.
Declared terms for API access, model weights, or self-hosting.
Limits and integration
| API ID | inclusionAI/invoker-render-1.0 |
|---|---|
| Model type | Video |
| Access model | Downloadable weights |
| License | Apache-2.0 + CC BY-NC 4.0 · non-commercial |
| Deployment | Self-hosted |
| Release | 16 NOV 2026 |
| Knowledge cutoff | Not published |
| Entrance | Text · Image · Video |
| Exit | Video |
| Context window | 1344 × 768 RGB image, 24 fps proxy video and text |
| maximum output | Streaming video; file inference with 17 × N + 5 frames, minimum 39 |
| Reasoning | Not published |
| Published tools | |
| Structured Outings | Not published |
| Batch processing | Not published |
| Prompt cache | Not published |
| Fine-tuning | Not published |
| Verified platforms | Hugging Face |
Research generation controlled by camera, geometry and motion. The model maintains a visual reference and a bounded history window. The project includes a WebSocket service and file and browser clients.
Use is subject to non-commercial terms. The 24 fps describe the timeline, not inference speed. The production profile uses eight NVIDIA GPUs; the main FP32 weights occupy about 132.5 GB. The world engine and proxies are external components.
What it can do
Orientation
Research generation controlled by camera, geometry and motion. The model maintains a visual reference and a bounded history window. The project includes a WebSocket service and file and browser clients.
Context
1344 × 768 RGB image, 24 fps proxy video and text · Streaming video; file inference with 17 × N + 5 frames, minimum 39
Tools and integration
Not published
Access
Hugging Face · downloadable checkpoint
Documented cost
| Concept | Worth | Unit/condition |
|---|---|---|
| Standard input | No published official API rate | Standard API rate |
| Cached input | Not published | Reading reused prefixes |
| Cache write or storage | Not published | The condition varies by provider |
| Standard output | Self-hosted infrastructure cost | May include reasoning tokens |
| Batch input | Not published | Asynchronous processing |
| Batch output | Not published | Asynchronous processing |
The card states Apache-2.0 together with CC BY-NC 4.0; retain the non-commercial restriction. No official API tariff is verified. Self-hosting costs include GPUs, memory, proxy generation and storage.
i Prices change and may depend on level, region or context length. Check the source before making a decision.
How to read the results
The card describes the input contract and four-step causal streaming. No comparable numeric benchmark is verified for this checkpoint. Session continuity does not prove real-time speed or unlimited stability.
i We have not found published benchmarks for this model.
Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.
Chronology
inclusionAI/invoker-render-1.0
Weights published on the 16th; contract and license verified in the November 18 revision.