LLM Reference

SubQ Models by Subquadratic

SubquadraticProprietary
1 model2026Up to 1m ctx

Last refreshed 2026-06-29. Next refresh: weekly.

Details

ResearcherSubquadratic
LicenseProprietary
Commercial useCommercial use: conditional
Models1
Released2026
Max context1m

Capabilities

ReasoningAll models
JSON / Tool useAll models

Links

Website

About

The SubQ model family from Subquadratic Inc. features a fully sub-quadratic sparse-attention architecture that scales linearly with context length (O(n) complexity). Designed for long-context reasoning, the family targets 12M-token context windows with significantly reduced compute costs vs. standard transformer-based models.

Decision facts

Best fit
reasoningJSON / Tool use
Capability starting point
SubQ 1M-Preview with 1m context and reasoning and JSON / Tool use
Lowest tracked input
Not tracked
Closest related family
Claude 3

Current Variants

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

1 in view

Use when the workload needs 1m context, reasoning, and JSON / Tool use.

2026-051m contextreasoningJSON / Tool use

Release Timeline

1 release group
2026-05
1 current
SubQ 1M-Preview
1m contextreasoningJSON / Tool use
Current

Specifications(1 models)

SubQ model specifications comparison
ModelReleasedContextReasoningJSON / Tool use
SubQ 1M-Preview2026-051mYesYes

Available From(1 provider)