← All updates

llm

Introducing Continuum1-9B

A fully linear 8.6B foundation model with 2M native context — hybrid GLA + Gated DeltaNet, open weights, and production kernels on Hugging Face.

Innomium LLM Team · March 1, 2026

Continuum1-9B is Innomium's long-context foundation model for reasoning, math, and document-scale workflows.

Highlights

  • ~8.6B parameters (BF16, sharded safetensors)
  • 2,097,152-token native context with linear compute
  • 75.0% MMLU and strong math benchmarks at a fraction of quadratic attention cost

Architecture

Hybrid Gated Linear Attention and Gated DeltaNet layers with No Positional Embeddings (NOPE) enable stable extrapolation far beyond the training window.

Load from innomium/Continuum1-9B with trust_remote_code=True and the Innomium kernel stack.

Want production AI shipped on your timeline?

Talk with Innomium about vision models, long-context LLMs, or builder challenges.

Schedule a call