Instructions to use MatterWaveAI/grt-mamba-b0.7.0-20m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MatterWaveAI/grt-mamba-b0.7.0-20m with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="MatterWaveAI/grt-mamba-b0.7.0-20m")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("MatterWaveAI/grt-mamba-b0.7.0-20m", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use MatterWaveAI/grt-mamba-b0.7.0-20m with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MatterWaveAI/grt-mamba-b0.7.0-20m"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MatterWaveAI/grt-mamba-b0.7.0-20m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/MatterWaveAI/grt-mamba-b0.7.0-20m

SGLang

How to use MatterWaveAI/grt-mamba-b0.7.0-20m with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "MatterWaveAI/grt-mamba-b0.7.0-20m" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MatterWaveAI/grt-mamba-b0.7.0-20m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "MatterWaveAI/grt-mamba-b0.7.0-20m" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MatterWaveAI/grt-mamba-b0.7.0-20m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use MatterWaveAI/grt-mamba-b0.7.0-20m with Docker Model Runner:
```
docker model run hf.co/MatterWaveAI/grt-mamba-b0.7.0-20m
```

GRT-Mamba-b0.7.0-20m (`MatterWaveAI/grt-mamba-b0.7.0-20m`)

GRT-Mamba-b0.7.0-20m is the 20M (b0.7.0-20m) Compact Purely Rotational Selective State-Space Model (Rotational-SSM), Unitary World Model & O(1) Constant-Memory Trajectory Judge (~26.4M stored parameters / ~66M dense-equivalent capacity, 0 dense linear matrices, 0 1D/2D convolutional layers, 0 bytes KV cache).

Built on MatterWaveAI/grt-20m as the 20M edge and scaling-ladder rung of GRT-Mamba, GRT-Mamba-b0.7.0-20m formulates selective state-space recurrence through norm-preserving orthogonal transformations (det = +1.000000):

Zero Vanishing or Exploding Gradients by Construction: Homogeneous state magnitude drifts by < 5.0e-7 over 1,000+ recursive steps (< 1e-14 at T = 8,192 in float64).
Parallel Training Scan + O(1) Recurrent Inference: Trains via a parallel prefix scan (< 3.82e-6 numerical parity with step-by-step recurrence) while executing inference in O(1) compute and O(1) memory per token (0 bytes of KV cache, 2.25 KB fixed recurrent state) across arbitrary sequence lengths (T = 128 -> 8,192+).

Release Note (b0.7.0 Preview Card): This card reports 20M (b0.7.0-20m) held-out XL-smol verification (12,003 held-out XL-smol records / 29,609 questions and 17,130 12-domain XL-smol test records / 44,287 questions), SLM-Bench & GSM8K reasoning benchmarks (13,644 examples), parallel-scan parity, and O(1) memory scaling. Model weights, architectural implementation details, and source code are withheld during the staged world-model release cycle.

`XL-smol` Benchmark Results (`GRT-Mamba-b0.7.0-20m`)

1. Rotational State-Space (`GRT-Mamba-b0.7.0-20m`) & Scan Invariants

Architectural Invariant / Benchmark	Target Gate	`GRT-Mamba-b0.7.0-20m` Result	Status
Dense Linear / `1D` / `2D` Convolutional Layers	`0 / 0 / 0`	`0 / 0 / 0` (100% Orthogonal)	PASS
Parallel Prefix Scan vs. `O(1)` Step Recurrence Parity	`< 1e-5`	`3.81e-6` (`Exact Match`)	PASS
Selective State Transition Determinant	`+1.000000`	`+1.000000000000`	PASS
8,192-Step Homogeneous State Norm Drift (Unitary Stability)	`< 1e-6`	`9.55e-15` f64 / `< 5.0e-7` f32	PASS
KV-Cache Memory Growth (`T = 128 -> 8,192`, Pure `SSM`)	`0` bytes	`0` bytes (`2.25 KB` fixed state)	PASS
Collapse-Free State Effective Rank Retention (`0` Negatives)	`>= 90.0%`	`>= 96.0%`	PASS

2. `GRT-Mamba-b0.7.0-20m`: `XL-smol` System-1 Rotational SSM World-Model Judge & MPC Router (`12,003` Held-Out Records / `29,609` Questions)

`GRT-Mamba-b0.7.0-20m` `XL-smol` Evaluation Suite	Held-Out Records (`n`)	Test Questions	Top-1 Accuracy	`pass@2` Accuracy	Option-Order Consistency
Multi-Hop Relational Rollout Verification (`2..3` Hops)	`2,500`	`5,000`	`90.40%`	`96.10%`	`100.0%` (`PASS`)
Causal Science & Physical Commonsense Verification	`2,500`	`5,000`	`95.30%`	`98.80%`	`100.0%` (`PASS`)
Single-Hop Factual World-Model Verification	`2,500`	`5,000`	`86.80%`	`93.10%`	`100.0%` (`PASS`)
4-Mode Socratic & Epistemic Verification (`SEE`/`SAY`/`DO`/`KNOW`)	`2,500`	`5,000`	`64.20%`	`92.80%`	`100.0%` (`PASS`)
2D Visual-Spatial Grid Rule Induction (`ARC-AGI-2` `XL-smol`)	`1,200`	`3,600` core (`7,200` all)	`56.17%`	`81.33%`	`100.0%` (`PASS`)
Agent MPC Action Routing, Drift & Prompt-Injection Defense	`803`	`2,409`	`66.75%`	`71.96%`	`100.0%` (`PASS`)
6-Suite `XL-smol` Aggregate (`Macro / Micro`)	`12,003`	`26,009` core (`29,609` all)	`76.60%` / `79.19%`	`89.87%`	`100.0%` (`PASS`)
Consolidated 12-Domain `XL-smol` Benchmark (`data_master_v9_xl`)	`17,130`	`40,687` core (`44,287` all)	`73.34%` Macro	`75.94%` Core Micro	`100.0%` (`PASS`)

3. `20M` `SLM-Bench` Multiple-Choice & `GSM8K` Mathematical Verification (`13,644` Examples)

Benchmark (`20M` Rung)	Split (`n`)	Random	`GRT-Mamba-b0.7.0-20m` (24h Curriculum)	`GRT-Mamba-b0.7.0-20m` (`10C` Horizon)
BoolQ (`google/boolq`)	`val` (`3,270`)	`50.00%`	`66.70%`	`73.45%`
QNLI (`nyu-mll/glue`)	`val` (`1,000`)	`50.00%`	`78.70%`	`86.58%`
ARC-Easy (`allenai/ai2_arc`)	`test` (`2,376`)	`25.00%`	`38.05%`	`55.76%`
ARC-Challenge (`allenai/ai2_arc`)	`test` (`1,172`)	`25.00%`	`30.55%`	`40.70%`
OpenBookQA (`allenai/openbookqa`)	`test` (`500`)	`25.00%`	`45.40%`	`55.04%`
PIQA (`baber/piqa`)	`val` (`1,838`)	`50.00%`	`57.40%`	`65.78%`
WinoGrande (`allenai/winogrande`)	`val` (`1,267`)	`50.00%`	`52.64%`	`63.26%`
HellaSwag (`Rowan/hellaswag`)	`val` (`1,000`)	`25.00%`	`37.80%`	`51.18%`
CommonsenseQA (`tau/commonsense_qa`)	`val` (`1,221`)	`20.00%`	`28.83%`	`41.58%`
9-Task Macro Average	`13,144` total	`35.56%`	`48.45%`	`59.26%`
GSM8K 4-Way Verification	`test` (`500`)	`25.00%`	`98.60% (PPL 7.51)`	`98.80% (PPL 2.14)`

Coming Next: Scaled Rotational State-Space Models (`GRT-Mamba-100M` & `GRT-Mamba-350M`)

Building directly on the 10M, 20M, and 30M GRT-Mamba and S1-GRT-Mamba scaling ladder, our next scaled releases expand O(1) constant-memory selective rotational state-space modeling and streaming System-1 MPC verification:

GRT-Mamba-100M (~195M Dense-Equivalent Capacity — Scaled Rotational SSM World Model & Streaming S1 Judge):
- Scales O(1) selective rotational state-space recurrence (grt-mamba) and streaming MPC trajectory verification (s1-grt-mamba) in a deep 0-linear-layer architecture—projected on our empirical scaling ladder to reach 64.9% SLM-Bench macro accuracy, 89.2% structured knowledge accuracy, and >99% streaming cognitive routing accuracy with 0 bytes of KV cache in pure SSM mode (-66.7% in Hybrid mode).
GRT-Mamba-350M (~920M Dense-Equivalent Capacity — Flagship Rotational State-Space Foundation Model):
- Scales long-horizon O(1) selective rotational SSM dynamics and multi-step MPC safety verification across arbitrary stream lengths (93.8% structured knowledge accuracy, 97.0% cognitive telemetry routing).

`MatterWaveAI` Rotational State-Space (`GRT-Mamba` / `SSM`) Scaling Ladder (`10M` / `20M` / `30M`)

Rotational Selective State-Space (GRT-Mamba) World Model: MatterWaveAI/grt-mamba (10M: grt-mamba-b0.7.0-10m | 20M: grt-mamba-b0.7.0-20m)
Streaming O(1) Rotational SSM World-Model Judge: MatterWaveAI/s1-grt-mamba (10M: s1-grt-mamba-b0.7.0-10m | 20M: s1-grt-mamba-b0.7.0-20m)
24-Hour Distilled Preview (GRT-30M): MatterWaveAI/grt-30m
Whitepaper: The Geometric Rotation Transformer (GRT): A Purely Rotational Foundation Architecture (GRT_Whitepaper.pdf)

License & Citation

Released under the Apache 2.0 license.

@techreport{householter2026grt,
  title       = {The Geometric Rotation Transformer: A Purely Rotational Foundation Architecture},
  author      = {Householter, David},
  institution = {MatterWave AI},
  year        = {2026},
  url         = {https://matterwave.ai}
}

Downloads last month: -; Downloads are not tracked for this model. How to track

Model tree for MatterWaveAI/grt-mamba-b0.7.0-20m

Base model

MatterWaveAI/grt-20m

Finetuned

(3)

this model

Finetunes

1 model

GRT-Mamba-b0.7.0-20m (MatterWaveAI/grt-mamba-b0.7.0-20m)

XL-smol Benchmark Results (GRT-Mamba-b0.7.0-20m)

1. Rotational State-Space (GRT-Mamba-b0.7.0-20m) & Scan Invariants

2. GRT-Mamba-b0.7.0-20m: XL-smol System-1 Rotational SSM World-Model Judge & MPC Router (12,003 Held-Out Records / 29,609 Questions)

3. 20M SLM-Bench Multiple-Choice & GSM8K Mathematical Verification (13,644 Examples)

Coming Next: Scaled Rotational State-Space Models (GRT-Mamba-100M & GRT-Mamba-350M)

MatterWaveAI Rotational State-Space (GRT-Mamba / SSM) Scaling Ladder (10M / 20M / 30M)