Octane 2027 Roadmap and Beyond
Posted: Fri Nov 28, 2025 7:43 am
Hi all,
We are incredibly excited to share a preview of Octane 2027 and our roadmap for the year ahead.
Here is a summary of the main features being actively developed in the next Octane 2027.1 release:
Real-time Neural Viewport rendering
Octane 2027 features a new real-time neural rendering mode, built on Octane 2026’s NRC framework, which enables near noise-free interactive rendering without any compromises to Octane’s spectral path-tracing, even in complex lighting conditions.
Here you can see a preview of the new real-time mode in a variety of environments, featuring camera movement and object and light manipulation in real-time while keeping viewport interactivity.
Render to Gaussian Splat
Octane 2027 will allow rendering directly to Neural Objects and Gaussian splat without intermediate rendering - an industry first. This will allow users to generate a splat file as the render output of the entire scene, and load it into applications with support for standard splat formats. This system can make use of spherical harmonics, allowing view-dependent effects to be properly stored in the output file.
Octane Splash Screen Rendered to Gaussian Splat in Octane 2027 (preview in SuperSplat):
Live Splat Renders of production scenes from Octane 2027:
Also previews at SuperSplat for Star Trek USS Enterprise Ship Exterior, bridge interiors, and Space Station.
Below is a sample image of an output Gaussian splat generated by Octane and then rendered back by Octane (preview at SuperSplat). This was created with over 14 million spherical harmonic samples. It took just 2 1/2 minutes to calculate on a single machine. The object itself is made up of over 500K splats.
Here is also an ORBX file containing the output PLY splat file as well as the original scene used to generate it.
Anime and Sketch Rendering
Octane’s toon rendering is getting a complete overhaul with a new system that allows rendering of the triangular mesh edges, which can be used to highlight silhouettes for stylized or anime-style rendering.
The edges can extend beyond the geometry bounds and their appearance can be fully configured, including width, tint color, texture and displacement.
The render below shows the toon edge feature applied to a complex 3d model of a robot using a custom stroke texture that is thin on the ends and wide in the middle.
The video below shows the toon edge applying realtime stylized silhouettes to a model of a simple city where the original model is also displayed for reference, then when the toon edge feature is on, the splining feature is toggled showing that the edges can be used to drive a spline of the edges. It also shows using a stroke texture that is like a scribble.
Generative PBR Materials
We are introducing a new generative text-to-image workflow that allows the generation of quality PBR materials using an artist defined text prompt which enables the creation of new materials almost immediately running locally on your machine. This can help iterating through different concepts much faster than before.
Improvements to light and environment sampling
As our continuous efforts in making Octane render more efficiently and noise free faster, we are working on two separate techniques that should help with this:
Environment map visibility cache
This is a method to improve HDRI environment sampling which helps especially under low light conditions and specially suited to render indoor environments where most of the light contribution is provided by an external source.
The image below shows the same interior scene, where the main light contribution comes from an external HDRI environment, rendered in Octane with and without using the new environment map visibility cache system. The reduction in noise is clearly noticeable, especially in those areas where there’s the least light contribution.
Even at one sample the difference is still clearly noticeable:
AI Light 2.0 for many light sampling
This is a hierarchical light-sampling algorithm that improves on earlier Octane’s AI Light and traditional power-based sampling by incorporating full spatial and directional information for all emissive primitives, which allows sampling the emitters with the most contribution, delivering more accurate results with significantly reduced noise.
Below is a comparison between the new method and Octane’s earlier sampling algorithm:
Amazon Lumberyard Bistro, Open Research Content Archive (ORCA) 2017
The difference is even more noticeable between the new method and Octane’s AI light sampling algorithm:
Amazon Lumberyard Bistro, Open Research Content Archive (ORCA) 2017
Procedural Volume Generation
We are working on adding support for generation of parametric procedural volumetric objects inside Octane such as clouds, smoke or foam.
Vectron to Mesh
This feature allows converting a Vectron object or an entire Vectron graph to a triangular mesh, one of the advantages of being able to do this is for it to be used to scatter other objects onto the procedural surface.
Wave Optics
New wave optics concepts such as birefringence and diffraction grating are coming to Octane to make your renders feel even more realistic.
The image below showcases several effects simulated with the new wave-optics universal material. On the left, the glass cube displays pure birefringence with its characteristic double-image effect. The glass sphere on the right carries a thin birefringent coating, producing interference patterns that arise from the combination of birefringence and thin-film effects. The compact disc, smartphone, and gift wrap exhibit diffraction-grating patterns caused by tiny periodic surface structures that split light into coloured interference bands. The paper-clip packaging also shows stress-induced birefringence, a common effect in strained plastics.
OTOY Studio Integration
Octane is also bringing you a new integration with the upcoming OTOY Studio service which will allow for new generative and augmentative AI workflows inside the Octane node graph. Some of these include text-to-image, image-to-image, image-to-mesh, text-to-video, etc. and will allow for faster iterations while developing new concepts and generation of new assets for your Octane projects.
4D Video editor in OTOY Studio:
You can see the live scene shown in the OTOY Studio 4D video editor here.
GPU Mesh Deformations
Octane currently supports two forms of geometry modifiers: bone deformations, and vertex displacement. These currently happen on the CPU, which limits the ability to edit this in real-time. We are planning to move part, if not all of this logic to the GPU, which should make editing these types of scenes much faster for each iteration.
Unlimited UV maps
Octane is currently limited to only 3 sets of UV maps per mesh. We are aware that with some workflows this can be a limitation especially with some specific DCC applications, our plan is to extend this to either a larger number or if possible, an unlimited number of UVs.
USD Export
Octane has had support for importing and rendering USD scenes for some time now. To continue with our support for this industry standard which we know is at the heart of many user workflows, we are now also working on allowing users to also export USD from Octane as an exchange format with other applications with support for it.
Python API
Octane has had support for an API using Lua since the early days, however we know many users are familiar with the Python programming language, which is at the same time incredibly powerful due to the amount of available libraries available for it. Because of this, we are working on a second API that will expose the same aspects of Octane, giving users a second option to programmatically interact with Octane and their Octane scenes.
Forward Compatibility for NVIDIA and Apple GPUs
Octane has usually been one of the first applications to add support for the latest hardware as soon as it’s been made available, however this process usually takes some time until a new version of Octane is released and needs to be done each time a new architecture is available, forcing updates that are often not ideal. We are hoping this won’t be the case for much longer, so newer versions of Octane should be able to work on future hardware versions that did not exist at the time the application was released, as long as the right driver is being installed with support for it.
Happy rendering,
Your OTOY Team
We are incredibly excited to share a preview of Octane 2027 and our roadmap for the year ahead.
Here is a summary of the main features being actively developed in the next Octane 2027.1 release:
- Real-time Neural IPR rendering
- Render to Gaussian Splat
- Anime and Sketch Renderer
- Diffraction and Wave optics
- Generative PBR Materials
- OTOY Studio integration
- Procedural volume, cloud, water, smoke generation
- AI Light 2.0
- Vectron to mesh
- Unlimited UV maps
- GPU mesh deformations
- Full scene USD export
- Python API
- Future Proof support for new Apple and NVIDIA GPUs
Real-time Neural Viewport rendering
Octane 2027 features a new real-time neural rendering mode, built on Octane 2026’s NRC framework, which enables near noise-free interactive rendering without any compromises to Octane’s spectral path-tracing, even in complex lighting conditions.
Here you can see a preview of the new real-time mode in a variety of environments, featuring camera movement and object and light manipulation in real-time while keeping viewport interactivity.
Render to Gaussian Splat
Octane 2027 will allow rendering directly to Neural Objects and Gaussian splat without intermediate rendering - an industry first. This will allow users to generate a splat file as the render output of the entire scene, and load it into applications with support for standard splat formats. This system can make use of spherical harmonics, allowing view-dependent effects to be properly stored in the output file.
Octane Splash Screen Rendered to Gaussian Splat in Octane 2027 (preview in SuperSplat):
Live Splat Renders of production scenes from Octane 2027:
Also previews at SuperSplat for Star Trek USS Enterprise Ship Exterior, bridge interiors, and Space Station.
Below is a sample image of an output Gaussian splat generated by Octane and then rendered back by Octane (preview at SuperSplat). This was created with over 14 million spherical harmonic samples. It took just 2 1/2 minutes to calculate on a single machine. The object itself is made up of over 500K splats.
Here is also an ORBX file containing the output PLY splat file as well as the original scene used to generate it.
Anime and Sketch Rendering
Octane’s toon rendering is getting a complete overhaul with a new system that allows rendering of the triangular mesh edges, which can be used to highlight silhouettes for stylized or anime-style rendering.
The edges can extend beyond the geometry bounds and their appearance can be fully configured, including width, tint color, texture and displacement.
The render below shows the toon edge feature applied to a complex 3d model of a robot using a custom stroke texture that is thin on the ends and wide in the middle.
The video below shows the toon edge applying realtime stylized silhouettes to a model of a simple city where the original model is also displayed for reference, then when the toon edge feature is on, the splining feature is toggled showing that the edges can be used to drive a spline of the edges. It also shows using a stroke texture that is like a scribble.
Generative PBR Materials
We are introducing a new generative text-to-image workflow that allows the generation of quality PBR materials using an artist defined text prompt which enables the creation of new materials almost immediately running locally on your machine. This can help iterating through different concepts much faster than before.
Improvements to light and environment sampling
As our continuous efforts in making Octane render more efficiently and noise free faster, we are working on two separate techniques that should help with this:
Environment map visibility cache
This is a method to improve HDRI environment sampling which helps especially under low light conditions and specially suited to render indoor environments where most of the light contribution is provided by an external source.
The image below shows the same interior scene, where the main light contribution comes from an external HDRI environment, rendered in Octane with and without using the new environment map visibility cache system. The reduction in noise is clearly noticeable, especially in those areas where there’s the least light contribution.
Even at one sample the difference is still clearly noticeable:
AI Light 2.0 for many light sampling
This is a hierarchical light-sampling algorithm that improves on earlier Octane’s AI Light and traditional power-based sampling by incorporating full spatial and directional information for all emissive primitives, which allows sampling the emitters with the most contribution, delivering more accurate results with significantly reduced noise.
Below is a comparison between the new method and Octane’s earlier sampling algorithm:
Amazon Lumberyard Bistro, Open Research Content Archive (ORCA) 2017
The difference is even more noticeable between the new method and Octane’s AI light sampling algorithm:
Amazon Lumberyard Bistro, Open Research Content Archive (ORCA) 2017
Procedural Volume Generation
We are working on adding support for generation of parametric procedural volumetric objects inside Octane such as clouds, smoke or foam.
Vectron to Mesh
This feature allows converting a Vectron object or an entire Vectron graph to a triangular mesh, one of the advantages of being able to do this is for it to be used to scatter other objects onto the procedural surface.
Wave Optics
New wave optics concepts such as birefringence and diffraction grating are coming to Octane to make your renders feel even more realistic.
The image below showcases several effects simulated with the new wave-optics universal material. On the left, the glass cube displays pure birefringence with its characteristic double-image effect. The glass sphere on the right carries a thin birefringent coating, producing interference patterns that arise from the combination of birefringence and thin-film effects. The compact disc, smartphone, and gift wrap exhibit diffraction-grating patterns caused by tiny periodic surface structures that split light into coloured interference bands. The paper-clip packaging also shows stress-induced birefringence, a common effect in strained plastics.
OTOY Studio Integration
Octane is also bringing you a new integration with the upcoming OTOY Studio service which will allow for new generative and augmentative AI workflows inside the Octane node graph. Some of these include text-to-image, image-to-image, image-to-mesh, text-to-video, etc. and will allow for faster iterations while developing new concepts and generation of new assets for your Octane projects.
4D Video editor in OTOY Studio:
You can see the live scene shown in the OTOY Studio 4D video editor here.
GPU Mesh Deformations
Octane currently supports two forms of geometry modifiers: bone deformations, and vertex displacement. These currently happen on the CPU, which limits the ability to edit this in real-time. We are planning to move part, if not all of this logic to the GPU, which should make editing these types of scenes much faster for each iteration.
Unlimited UV maps
Octane is currently limited to only 3 sets of UV maps per mesh. We are aware that with some workflows this can be a limitation especially with some specific DCC applications, our plan is to extend this to either a larger number or if possible, an unlimited number of UVs.
USD Export
Octane has had support for importing and rendering USD scenes for some time now. To continue with our support for this industry standard which we know is at the heart of many user workflows, we are now also working on allowing users to also export USD from Octane as an exchange format with other applications with support for it.
Python API
Octane has had support for an API using Lua since the early days, however we know many users are familiar with the Python programming language, which is at the same time incredibly powerful due to the amount of available libraries available for it. Because of this, we are working on a second API that will expose the same aspects of Octane, giving users a second option to programmatically interact with Octane and their Octane scenes.
Forward Compatibility for NVIDIA and Apple GPUs
Octane has usually been one of the first applications to add support for the latest hardware as soon as it’s been made available, however this process usually takes some time until a new version of Octane is released and needs to be done each time a new architecture is available, forcing updates that are often not ideal. We are hoping this won’t be the case for much longer, so newer versions of Octane should be able to work on future hardware versions that did not exist at the time the application was released, as long as the right driver is being installed with support for it.
Happy rendering,
Your OTOY Team
Wow!! Amazing!!!!! 
