# SenseTime — SenseNova Loop-9B正式版：让更少的参数走更深的计算路径

- Company: SenseTime (sensetime.com)
- Announced: 2026-11-17T16:00:06.805000+00:00
- Category: new-model
- Coverage: not counted
- Announcement: yes
- Group: models
- Source: https://www.sensetime.com/cn/news/sensenova-loop-9b-20261118-1059
- Record: https://forck.live/items/37773-sensenova-loop-9b
- Title in English (translated by forck.live from zh): SenseNova Loop-9B official release: fewer parameters, deeper compute paths
- Subject: SenseNova / 商汤
- Open weights: yes
- Models affected: SenseNova Loop-9B
- Context window: 512K tokens

SenseTime has released SenseNova Loop-9B, a roughly 9B-parameter multimodal model that reuses shared decoder layers to reach the performance of much larger models. The model runs its shared layers twice, turning 32 physical decoder layers into 56 layer computations per forward pass, and pairs that loop with dual-stream manifold-constrained hyper-connections. SenseTime says the design keeps deployment cheap: 4bit weights take about 4.9GB and 256K-context inference needs roughly 15 GiB of memory. On Agent and data-analysis tests the company reports NL2Repo at 41.36 against 44.60 for Qwen3.8-27B, and AIDABench at 65.31 against 65.01. An official AWQ 4bit quantised version was offered on day one, and the weights and code are open for download. SenseTime also reports that raising loop counts from 2 to 4 still improved LiveCodeBench V6 and AIDAbench-data_visualization scores, and says it will study a dynamic loop architecture that adapts loop count to task complexity.

## Evidence

Verbatim from https://www.sensetime.com/cn/news/sensenova-loop-9b-20261118-1059:

> SenseTime officially released SenseNova Loop-9B. This model continues the Preview version's exploration of parameter efficiency: by reusing model layer parameters through the Loop structure and combining it with manifold-constrained hyper-connections (mHC), it expands computational depth at approximately 9B parameters, targeting complex reasoning, code, tool calling, and multimodal tasks. Benefiting from the efficient loop structure, the official 4-bit quantized weights are only about 4.9GB, with deployment memory requirements as low as under 10GB, significantly broadening the range of deployable hardware and enabling low-memory chips to achieve performance comparable to cutting-edge 27B-level models.
> 
> Since the Preview release, we have received community feedback and attention regarding model usage, inference deployment, and ecosystem adaptation. This official version retains the Loop and dual-stream mHC architecture, further optimizing the model's performance in Agent, code, and data analysis tasks, while continuing to explore the role of recurrent computation in enhancing model capabilities.
> 
> SenseNova Loop-9B has approximately 9B total parameters. By executing the shared intermediate layers twice, the model expands 32 physical decoder layers into 56 layer computations; it supports text and image input with text output, and a context length of up to 512K tokens. In Agent and data analysis evaluations, the model demonstrates competitiveness comparable to Qwen3.8-27B: an NL2Repo score of 41.36, close to the latter's 44.60; and an AIDABench score of 65.31, higher than the latter's 65.01.
> 
> Model Highlights
> 
> Shared Recurrent Module: Layers 4–27 form a shared-weight recurrent module executed twice, enabling the model to complete 56 decoder layer computations in a single forward pass.
> 
> Dual-Stream mHC: Information routing and fusion of two residual streams in the recurrent module via learnable mappings.
> 
> Hybrid Attention: Decoders are grouped in sets of four layers, using a configuration of three layers of linear attention and one layer of full attention.
> 
> Long-Context Multimodal: Training covers text, images, code, and tool call trajectories, with context length extended up to 512K tokens.
> 
> Low-Resource Private Deployment Friendly: With 4-bit weight quantization and 8-bit KV cache, memory requirements for 256K long-context inference are approximately 15 GiB, helping lower the barrier for local and private deployment.
> 
> Official Quantized Version Available Day 0: The official AWQ 4-bit quantized version is provided on the first day of release for quick deployment.
> 
> SenseNova Loop-9B has released model weights and related code for developers and researchers to download, deploy, and explore.
> 
> Model Capabilities: Official Version Further Improved Over Preview
> 
> We evaluated the official SenseNova Loop-9B on tasks including general Agent, code, and data analysis, comparing it with the Preview version and models like Qwen3.8-27B.
> 
> Evaluation results show that the official SenseNova Loop-9B has significantly improved over the Preview version on multiple benchmarks; at approximately 9B parameters, it demonstrates performance on par with Qwen3.8-27B.
> 
> Model Performance

---

Record: https://forck.live/items/37773-sensenova-loop-9b
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md
