SubQ Snap Models by Subquadratic
Subquadratic releases · 2 in the last 12 months · this family litChangelog →
1 model2026Up to 384K ctxFrom $0.050/1M input
Last refreshed 2026-10-13. Next refresh: weekly.
Details
ResearcherSubquadratic
LicenseProprietary
Commercial useCommercial use: conditional
Models1
Released2026
Max context384K
Capabilities
VisionAll models
MultimodalAll models
Structured OutputsAll models
Links
WebsiteAbout
SubQ Snap is Subquadratic's decision-model family. A decision model reads a state (text, JSON, or images) plus named noul, choice, or score questions and returns probabilities over the allowed answers instead of generated text. Snap uses Subquadratic Sparse Attention (SSA) with a decision head and is served on the SubQ Decisions API (POST /v1/systemone). Not a chat LLM family; the SubQ language models are a separate family.
Decision facts
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
1 in view
SubQ SnapCurrent
Use when the workload needs 384K context, structured outputs, and multimodal inputs.
2026-10384K contextstructured outputsmultimodal inputs
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| SubQ Snap | Use when the workload needs 384K context, structured outputs, and multimodal inputs. | 2026-10 | 384K contextstructured outputsmultimodal inputs | Current |
Release Timeline
1 release group2026-10
1 current
SubQ Snap
Current384K contextstructured outputsmultimodal inputs
Specifications(1 models)
| Model | Released | Context | Vision | Multimodal | Structured Outputs |
|---|---|---|---|---|---|
| SubQ Snap | 2026-10 | 384K | Yes | Yes | Yes |