LLM Reference

XOR Models by Xyne

XyneApache 2.0Open sourceAgents
1 model2026Up to 262k ctx

Last refreshed 2026-09-24. Next refresh: weekly.

Details

ResearcherXyne
LicenseApache 2.0OSI-approved
Commercial useCommercial use: permitted
Models1
Released2026
Max context262k

Capabilities

VisionAll models
MultimodalAll models
Structured OutputsAll models

About

XOR is Xyne's open multimodal decision model family, a post-trained Qwen3.6 mixture-of-experts model for typed classification (choice, score, and noul) over text and images through a TypeSafe-compatible /v1/systemone API. Apache 2.0 weights are on Hugging Face at juspay/xor. The Hugging Face author is juspay; Xyne markets the release as xyne by juspay.

Decision facts

Best fit
agentsagentvision and multimodal work
Capability starting point
XOR with 262k context and structured outputs and multimodal inputs
Lowest tracked input
Not tracked
Closest related family
Agents-A1

Current Variants

Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.

1 in view
XORCurrent

Use when the workload needs 262k context, 35B parameters, and structured outputs.

2026-09262k context35B parametersstructured outputs

Release Timeline

1 release group
2026-09
1 current
XOR
262k context35B parametersstructured outputs
Current

Specifications(1 models)

XOR model specifications comparison
ModelReleasedContextParametersVisionMultimodalStructured Outputs
XOR2026-09262k35BYesYesYes