SenseNova Loop 9B Preview — Hardware Requirements & GPU Compatibility
VisionChatFunctions
SenseNova-Loop-9B-Preview is a preview multimodal language model from SenseNova for long-context and agentic workflows, accepting text and images and generating text. It is a looped transformer: a block of 24 decoder layers is run twice with shared weights, giving 56 layer executions from 32 distinct layers, with two-stream manifold-constrained hyper-connections routing information between residual streams. That gives a deeper execution path without duplicating the repeated block's parameters, so it has about 9 billion parameters but roughly 14.6 billion of inference compute. Context length is 262,144 tokens, and the model was released on October 30, 2026. It uses the Apache 2.0 license, which permits commercial use, and requires custom code to load. A 9B model should run on a single consumer GPU once quantized, though the repeated block adds compute over a standard 9B model.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 8K)Best on 24 GB (Q4_K_M · ~103 tok/s · 128K)Macs from 16 GB (Q4_K_M · ~22 tok/s)
Apache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run SenseNova Loop 9B Preview
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
ollama run hf.co/AtomicChat/SenseNova-Loop-9B-Preview-GGUF:Q4_K_MQ4_K_M fits fully on the RTX 4090 · 15.8 GB free · 16K context
Download: SenseNova-Loop-9B-Preview-Q4_K_M.gguf5.63 GBAtomicChat
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3080 (10 GB) (1 GB headroom) · Q4_K_M
- Generation speed
- ~78 tok/s
- generation speed
- Cost per 1M output tokens
- $0.40
- per 1M output tokens
How Much VRAM Does SenseNova Loop 9B Preview Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 4.7 GB | 8.7 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 128K context · ~140 tok/s | |
| 5.3 GB | 9.3 GB | 3-bit medium — 3-bit medium quantization | Fits: 128K context · ~124 tok/s | |
| 5.4 GB | 9.4 GB | 4-bit · legacy — 4-bit legacy quantization | Fits: 128K context · ~121 tok/s | |
| 6.4 GB | 10.4 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 128K context · ~103 tok/s | |
| 7.5 GB | 11.5 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 64K context · ~88 tok/s | |
| 8.5 GB | 12.6 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 64K context · ~77 tok/s | |
| 10.2 GB | 14.3 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 64K context · ~64 tok/s | |
| 19.9 GB | 23.9 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: 32K context · ~33 tok/s |
Selected Q4_K_M
4-bit medium quantization8.2GBat 16K context
- Weights
- 5.8 GB
- KV cache
- 2.2 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 15.8 GB free · ~103 tok/s
≥50 tok/s needs ≥489 GB/s · RTX 5090 ~183, RTX 3090 Ti ~103
Which Hardware Can Run SenseNova Loop 9B Preview?
Q4_K_M · 6.4 GBSenseNova Loop 9B Preview (Q4_K_M) requires 6.4 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 9+ GB is recommended. Using the full 262K context window can add up to 34.1 GB, bringing total usage to 40.5 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3070 Ti. 151 devices with unified memory can run SenseNova Loop 9B Preview, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), NVIDIA Jetson Orin Nano 8GB (Super).
12 GB
12 GB tier →fits · 3.8 GB free · up to 32K ctx
10 GPUs · 1 device family
RTX 3080 Ti~93Simulate output on RTX 3080 TiRTX 5070~69Simulate output on RTX 5070RTX 4070~52Simulate output on RTX 4070iPhone~9Simulate output on Apple iPhone 17 Pro20 GB
fits · 11.8 GB free · up to 64K ctx
1 GPU
RX 7900 XT~76Simulate output on RX 7900 XT32–40 GB
fits · 23.8 GB free · up to 128K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~183Simulate output on RTX 5090M5 Max36 GB~51Simulate output on Mac Studio M5 Max (36 GB)M532 GB~17Simulate output on MacBook Air 13" M5 (32 GB)Orin~21Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~14Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3.5 9BAlibaba | 99.4 | 6.6 GB | ~103 | same size |
| Gemma 2 9B ITGoogle | 68.0 | 7.3 GB | ~107 | same size |
| Ornith 1.5 9Bornith-ai | — | 6.6 GB | ~103 | same size |
| MiniCPM O 4 5OpenBMB | — | 8.3 GB | ~105 | same size |
| GLM 4.1V 9B ThinkingZ.ai | — | 7.2 GB | ~100 | same size |
| Shieldgemma 9BGoogle | — | 8.8 GB | ~107 | same size |
Specifications
- Publisher
- sensenova
- Parameters
- 9.7B
- Architecture
- SenseNovaLoopForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-10-28
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does SenseNova Loop 9B Preview need?
SenseNova Loop 9B Preview requires 6.4 GB of VRAM at Q4_K_M, or 19.9 GB at BF16. Full 262K context adds up to 34.1 GB (40.5 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 9.7B × 4.8 bits ÷ 8 = 5.8 GB
KV Cache + Overhead ≈ 0.6 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 34.7 GB (at full 262K context)
VRAM usage by quantization
What's the best quantization for SenseNova Loop 9B Preview?
For SenseNova Loop 9B Preview, Q4_K_M (6.4 GB) offers the best balance of quality and VRAM usage. Q5_K_M (7.5 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 4.7 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run SenseNova Loop 9B Preview on a Mac?
Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run SenseNova Loop 9B Preview. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 6.4 GB of usable unified memory (RAM minus macOS overhead).
Can I run SenseNova Loop 9B Preview locally?
Yes — SenseNova Loop 9B Preview can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is SenseNova Loop 9B Preview?
At Q4_K_M, SenseNova Loop 9B Preview can reach ~755 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~103 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 6.36 × 0.65 = ~818 tok/s
Estimated speed at Q4_K_M (6.4 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of SenseNova Loop 9B Preview?
At Q4_K_M, the download is about 5.79 GB. The full-precision BF16 version is 19.31 GB. The smallest option (Q2_K) is 4.10 GB.
Which GPUs can run SenseNova Loop 9B Preview?
52 consumer GPUs can run SenseNova Loop 9B Preview at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 7600. 40 GPUs have plenty of headroom for comfortable inference.
Which devices can run SenseNova Loop 9B Preview?
151 devices with unified memory can run SenseNova Loop 9B Preview at Q4_K_M (6.4 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.