tokenspeed
tokenspeed is 1️⃣拉 Docker 镜像并进容器：docker pull lightseekorg/tokenspeed-runner:latest，按官方文档启动容器后 git clone. It is ranked #597 on the Dev Radar, in AI dev tools, first seen 50 days ago and shared in 18 posts (361.9k views).
What tokenspeed says about itself
Recognize gfx1201 as RDNA4 so supported devices can use the existing portable Triton kernels and ROCm library implementations instead of failing platform detection. Keep RDNA4 out of the CDNA4-or-n...
Add gfx1201 support through portable AMD kernels by powderluv · Pull Request #1922 · lightseekorg/tokenspeed · GitHub Skip to content Navigation Menu Sign in Search / Sign in Sign up Appearance settings You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session. Dismiss alert {{ message }} Repository navigation Add gfx1201 support through portable AMD kernels - # 1922 # 1922 Merged zhyncs merged 1 commit into lightseekorg:main lightseekorg/tokenspeed:main from…
What people said about tokenspeed on X
Congrats to @lightseekorg on the impressive TokenSpeed results! 🚀 Great to see Qwen3.8-Flash-Next reaching 685 output tokens/s per user on Nvida GPUs. Excited to see the community unlock more of Qwen’s inference potential!
— @Alibaba_Qwen, 20 days ago · 997 likes · see the post
🫨我靠！Qwen3.8-Flash-Next 单用户推理速度直接干到 685 Token/s！谁还不知道？！ 阿里官方都来恭喜了，LightSeek 的 TokenSpeed 在 4× NVIDIA GB300 + NVFP4 上把这模型跑出了离谱速度。 主要亮点 🔹685.9 tokens/s per user（batch=1，SWE-Smith 基准），MTP 接受长度稳定在 ~3.23 🔹模型本身就是 Qwen4 架构预览：125B 总参数、仅 6B 激活、多模态 MoE + 51B N-gram Embedding 🔹优化硬核：QSA 稀疏注意力（微块选择）、内核融合、PDL 重叠、MTP 选择复用、PLE…
— @NFT_Chen, 20 days ago · 104 likes · see the post
Thanks to @lightseekorg for Day-0 support with TokenSpeed!
— @MiniMax_AI, 11 days ago · 188 likes · see the post
Alternatives to tokenspeed
- niklaslenz.com — Play the world in your browser. Every number, the exact command and the raw output
- goo.gle — Officially in beta, the AppFunctions API is opening up powerful new ways to integrate your functionality.
- dsh-superagent-prototype — DeepSeek open-sourced the design behind its always-on 'Super-agent': no chat window, and reportedly nearly half the…
- Agents in flight — Give it a try
- strata — and more
- flashprove — Further progress in provable inference will likely be an important part of improving trust in AI deployment. To that…
tokenspeed in numbers
- Rank on the Dev Radar: #597 of 9203
- Shared in 18 posts by 5 accounts: @Alibaba_Qwen, @NFT_Chen, @MiniMax_AI, @lightseekorg
- 361.9k views on those posts
- First seen 50 days ago, last shared 8 days ago
- Pricing seen by AGK Intelligence: open source
- Market: AI dev tools
FAQ
What is tokenspeed?
1️⃣拉 Docker 镜像并进容器：docker pull lightseekorg/tokenspeed-runner:latest，按官方文档启动容器后 git clone It was first shared on X 50 days ago and is ranked #597 on the Dev Radar.
Is tokenspeed free?
It is open source.
Who shared tokenspeed?
5 accounts on X, including @Alibaba_Qwen, @NFT_Chen, @MiniMax_AI, in 18 posts totalling 361.9k views.
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. 31.3k posts from 5k X accounts over the last 21 days, 9.2k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-03 23:11 UTC. Full method.