cuda.fast
cuda.fast is is LIVE. It is ranked #613 on the Dev Radar, in AI dev tools, first seen 9 days ago and shared in 3 posts (19.9k views).
What cuda.fast says about itself
Make Qwen 3.8 Flash Next decode faster on CUDA. Follow the official records and join the shared Yukon challenge.
cuda.fast — Qwen 3.8 Flash Next, faster with speculative decode Yukon cuda.fast All challenges Leaderboard Qwen 3.8 Flash Next runs 176.4% 0.0% 176.4% faster on DGX Spark . How the score works The official score is the single-stream composite: prefill gain to the 0.25 power times decode gain to the 0.75 power, so decode carries three quarters of the weight. A serial run is 1.0, and everything above it is real speculative-decode gain. benchd computes the score and seals it. cuda.fast @ 176.4% mlx.fast @ — How it works Participate Record By model Lin Log Improvement History Leaderboard current record 176.4% composite speed increase over the…
What people said about cuda.fast on X
MLX and Qwen 3.8 Next are getting some love, and they are using one of my flavours 😎 https://huggingface.co/Vontra/Qwen3.8-Flash-Next-MLX-4bit-MTP I have been tuning this since early this week and am currently sitting at 91.24 tok/s repeatedly, but at ~3.4k context with greedy sampling and a 512-token thinking cap.…
— @ashxhart, 9 days ago · 54 likes · see the post
https://cuda.fast is LIVE Apple M5 vs Nvidia DGX Spark. MLX vs CUDA. Same model, same graph. Let’s see how fast you can make Qwen 3.8-Flash-Next.
— @yukonresearch, 9 days ago · 41 likes · see the post
http://cuda.fast is LIVE Our first challenge: Qwen 3.8 Flash Next on NVIDIA and Apple, scored on one graph. 1. http://cuda.fast: 1 DGX Spark, starting at 17 to 18 tokens per second. 2. http://mlx.fast: 1 128 GB M5 MacBook, starting at 35 to 36. Submissions close September 24. Thanks to @TheDavidTai and @GumbiiDigital…
— @eigenlabs, 9 days ago · 34 likes · see the post
Alternatives to cuda.fast
- muse.ai — New connectors are live today. Come build with us.
- classifier.dev — now outperforms jev and is free
- mimo-v2.6 RL — Streaming the run
- academy.dair.ai — Chat with Paper
- Union Alpha — Union Alpha is a multimodal model built for research, coding, and agentic workflows, while delivering frontier-level…
- cua — Draft #3943
cuda.fast in numbers
- Rank on the Dev Radar: #613 of 1694
- Shared in 3 posts by 3 accounts: @ashxhart, @yukonresearch, @eigenlabs
- 19.9k views on those posts
- First seen 9 days ago, last shared 9 days ago
- Pricing seen by Jev: not stated
- Market: AI dev tools
FAQ
What is cuda.fast?
is LIVE It was first shared on X 9 days ago and is ranked #613 on the Dev Radar.
Is cuda.fast free?
Pricing is not stated on the page we read.
Who shared cuda.fast?
3 accounts on X, including @ashxhart, @yukonresearch, @eigenlabs, in 3 posts totalling 19.9k views.
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 15.1k posts from 4.7k X accounts over the last 21 days, 1.7k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-19 22:52 UTC. Full method.