Ripple-65B Preview, a dense 65B: first token at 128K takes 272.5 s on 1 DGX Spark and…
This is a AI post classified by AGK Intelligence as AI infra & evals (a model release), kept by the AI Radar because it carries real work, not commentary.
Ripple-65B Preview, a dense 65B: first token at 128K takes 272.5 s on 1 DGX Spark and 193.7 s on 2 1 DGX Spark, 128 GB: NVFP4 + DFlash2 Decode, code: 9.8 tok/s Decode, prose: 9.8 tok/s First token at 8K: 13.8 s First token at 128K: 272.5 s 8 requests at once: 10.9 tok/s total 2 DGX Sparks, 256 GB: BF16 + DFlash2 Decode, code: 9.2 tok/s Decode, prose: 11.0 tok/s First token at 8K: 10.1 s First token at 128K: 193.7 s 8 requests at once: 53.0 tok/s total Quant: 1 Spark NVFP4 on the MLP layers, attention in BF16, about 7.1 bpw overall. 2 Sparks full BF16 Setup, 1 Spark: the authors' Spark recip
Posted by Niklas (2.8k followers) 11 h ago · 4 likes · 622 views · view the original post on X. Kept by the AI Radar as AI infra & evals. Tools mentioned: serve.py, niklaslenz.com.
More AI work like this
- Enterprise AI is only as useful as the data and business context behind it. — @databricks
- DecodeWall: another benchmark I'm building, focused on concurrent calls. — @ivanfioravanti
- KindlingAI's GLM 5.3 Flash for one Spark recipe is here! 700k total KV on DGX OS (up to… — @mmastrac
- A small model coaching a giant? Aiona checks the held-out tie and SMF’s memory-card… — @MichaelGannotti
- Ranking inference providers six ways — @AdamHoltererer
- Harness adapters training is now live! — @yb2698
- NVIDIA is giving Kimi K3.1 Pro away for free. — @VaibhavSisinty
- The world’s first co-packaged optics Ethernet switch is in full production. — @NVIDIAAIInfra
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 65.6k posts from 5k X accounts over the last 14 days, 10.4k tools, 19 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 16:05 UTC. Full method.