Subscribe

SenseNova-Loop-9B

SenseTime released the final SenseNova-Loop-9B, a roughly 9B looped multimodal model with a 512K context, as open weights and code, following the October preview.

The final checkpoint replaces the October preview with updated weights and a longer 512K-token context. It keeps 32 decoder layers and runs the shared middle block (layers 4 to 27) twice, giving 56 layer computations per forward pass, with two residual streams using mHC. In SenseTime's own tests GAIA rises from 37.86 for the preview to 69.90, and Terminal-Bench 2.1 from 40.45 to 59.55. Cambricon reported launch-day support on its MLU platform.

Date
Tuesday 17 November 2026
Lab
SenseTime
Kind
open-weights
Access
open weights

Figures

MeasureValueMeasured by
GAIA69.90
37.86 for the Preview
company
Terminal-Bench 2.159.55
40.45 for the Preview
company
AIDABench65.31
Qwen3.8-27B scores 65.01
company
NL2Repo41.36
Qwen3.8-27B scores 44.60
company
Context window512K tokenscompany
4-bit AWQ weights sizeabout 4.9 GB
about 15 GiB GPU memory for 256K context with 8-bit KV cache, per model documentation
company

Benchmarks are developer-reported and not independently reproduced. Pandaily says the project is under Apache 2.0, and deployment docs still carry Preview-derived configurations.

Sources

  1. pandaily.com/sensetime-sensenova-loop-9b-final-cambricon-mlu-support
  2. technode.com/2026/11/18/sensetime-releases-open-weight-loop-9b-model-that-reuses-layers-fo

This record was checked against its sources on 18 November 2026. How we check

Read the daily brief for 17 November 2026