Sign inSign up
ripple-preview logo

ai/ripple-preview

Verified Publisher

By Docker

•Updated 2 days ago

Artifact
0

388

ai/ripple-preview repository overview

⁠Ripple-65B Preview NVFP4 (Experimental)

Ricursive Logo

BF16⁠ | NVFP4⁠ | Base Model⁠ | DFlash2⁠ | Launch Blog⁠ | License: Apache 2.0⁠

Ricursive Intelligence introduces Ripple-65B Preview, a 65B parameter dense Transformer model pretrained on 12 trillion tokens with support for up to 1M context length. Ripple-65B Preview delivers strong results on agentic coding, scoring 80.52 on TerminalBench 2.1 and surpassing significantly-larger US open-weight models such as Reflection Beam 501B (80.1), Meta Muse Spark Open 816B (76.4), Thinking Machines Inkling 975B (63.8), and NVIDIA Nemotron 3 Ultra 550B (56.4).

Our Preview release includes an instruction-tuned model in BF16 and NVFP4 numerics, and a BF16 base model. This repo contains the NVFP4 instruction-tuned checkpoint.

This is a conservative quantization of the BF16 checkpoint, using NVFP4 for the MLP layers while keeping attention in BF16.

⁠Run on DGX Spark

inference/local/ serves this checkpoint with the DFlash2 draft⁠ on a single DGX Spark (GB10, 128 GB unified memory) through stock, unmodified SGLang 0.5.19. You need Python 3.12, the CUDA 13 toolkit, build-essential and python3.12-dev.

hf download ricursiveAI/Ripple-65B-Preview-NVFP4 --include "inference/local/*" --include chat_template.jinja --local-dir ripple
cd ripple/inference/local
python3.12 -m venv .venv && .venv/bin/pip install -r requirements.txt
CUDA_HOME=/usr/local/cuda-13.0 PATH="$PWD/.venv/bin:/usr/local/cuda-13.0/bin:$PATH" python spark.py

spark.py downloads the checkpoints and starts SGLang's OpenAI-compatible server at http://127.0.0.1:8000/v1 with model name ripple. It supports streaming, tool calls and reasoning_content. ripple.ipynb uses the same engine from a notebook. In Python scripts, create Spark(...) under if __name__ == "__main__":, because SGLang starts worker processes.

curl http://127.0.0.1:8000/v1/chat/completions -H 'Content-Type: application/json' \
  -d '{"model":"ripple","messages":[{"role":"user","content":"Write a Python binary search."}],"max_tokens":256}'

Defaults:

  • Decoding: speculative tree decoding with 32 verification nodes from a 16-token draft, using the bundled kernels and deterministic Triton attention. Tree decoding is experimental. --decoding chain --kernels stock selects stock SGLang chain speculation.
  • Context: 131,072 tokens (input up to 131,065), with prefix caching on, so follow-up turns reuse their cached prefix.
  • Concurrency: up to 4 concurrent requests.
  • Sampling: temperature 1.0 and top-p 0.95 when a request omits them. Send "temperature": 0 for greedy output.

On a DGX Spark at temperature 1.0, this reaches about 19.5 tok/s for one request and 48 tok/s in total with four, on representative prompts with 128-token outputs.

⁠Ripple-65B-Preview: NVFP4 vs BF16 benchmarks

BenchmarkRipple-65B
NVFP4
Ripple-65B
BF16
Nemotron 3
Ultra 550B BF16
Inkling
975B BF16
Muse Spark
Open 816B BF16
Reflection
Beam 501B BF16
Model
ArchitectureDenseDenseMoEMoEMoEMoE
Total params65B65B550B975B816B501B
Active params65B65B55B41B42B23B
Pre-training tokens12T12T20T45TN/A23.8T
Knowledge
MMLU89.289.692.091.392.692.4
MMLU-Pro83.784.286.887.387.976.5
GPQA Diamond83.385.287.087.289.090.6
Math
AIME 202497.197.194.297.597.996.7
AIME 202597.998.394.291.7100.099.6
AIME 202695.895.095.497.198.397.7
HMMT Feb 202689.490.580.386.393.993.9
Coding & Agentic
Terminal-Bench 2.1 (self-reported)76.080.556.463.876.480.1
Terminal-Bench 2.1 (Terminus 2)76.080.555.156.278.770.8
SWE-bench Verified76.576.571.977.678.781.6
SWE-bench Pro v1.061.0661.4246.954.358.266.1
SWE-bench Pro v2.0 Hard77.879.130.739.672.476.5
DeepSWE 1.140.344.30.21.860.043.3
SciCode (Inspect AI)39.539.444.646.158.549.2
τ³-bench banking29.9031.2722.623.714.436.1
AutomationBench (Public)36.336.48.414.642.439.3
MCP Atlas73.2771.6762.276.089.280.8
Safety
FORTRESS Adversarial83.283.879.378.094.292.1
FORTRESS Benign95.595.589.195.982.686.4
StrongREJECT95.997.598.298.694.795.8
Instruction Following
IFBench78.679.281.779.878.679.6
Long Context
AA-LCR77.375.065.478.380.778.2

⁠Citation

@misc{ricursive2026ripple65b,
      title={Ripple-65B: Pushing the Frontier of Local Intelligence},
      author={Ricursive Intelligence},
      year={2026},
}

⁠Contact

For questions and comments, please contact us at [email protected]⁠. To learn more, please visit our website⁠.

Tag summary

Content type

Unrecognized

Digest

sha256:57ed3c42b…

Size

53.3 GB

Last updated

2 days ago

docker pull ai/ripple-preview

This week's pulls

Pulls:

140

Last week