superseedai.com mapple is available.
We introduce Maple-Preview, an open-source 20B-A1B ternary-weight reasoning LLM, SOTA in its weight class.
It solves IMO-level problems and runs at 200+ tokens/s on a Mac Mini M4, 5–16× faster than efficient models like Gemma 4, Qwen3.5, and gpt-oss.
deepgrove.ai/maple-preview
Leandro Barbosa retweeted
We introduce Maple-Preview, an open-source 20B-A1B ternary-weight reasoning LLM, SOTA in its weight class.
It solves IMO-level problems and runs at 200+ tokens/s on a Mac Mini M4, 5–16× faster than efficient models like Gemma 4, Qwen3.5, and gpt-oss.
deepgrove.ai/maple-preview
superseedai.com a new more model available thanks @PrismML
Today, we’re announcing Ternary Bonsai 2 27B.
Based on Qwen3.8 27B, Bonsai 2 27B is 9x smaller than its full-precision counterpart while retaining 98.2% of its aggregate benchmark performance.
Two months after the first Bonsai 27B release, the biggest change is quality. The footprint remains 5.9 GB, but the gap to full precision has narrowed materially, with particularly strong gains in agentic coding, multimodal reasoning, and long-horizon tool use.
Ternary Bonsai 2 27B is available today under Apache 2.0.
18
Leandro Barbosa retweeted
Pls avoid upgrading your DGX Spark to driver 610.43.02 / kernel 7.0. Some users report problems.
Failing kernel:
7.0.0-1019-nvidia
package: 7.0.0-1019.19~24.04.2
Thanks to @k1ng_pr4wn for reporting it.
forums.developer.nvidia.com/…
Leandro Barbosa retweeted
🚀 Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient.
🔹 Introducing the smallest model in our new architecture family, with native visual understanding.
🔹 Designed for greater capability, faster inference, higher throughput, and scaling to larger models.
1/6
Leandro Barbosa retweeted
For a long time you had to apologize for the quality of local models.
Now Local AI is actually good.
It turns out local hardware is already super capable - it was a model/software problem all along.
Leandro Barbosa retweeted
Yes, we are back👑, with 206 tok/s on a single RTX 5090!
Amazing Day-0 work from the SGLang team. Give it a try~@sgl_project
The king of small models is back! Qwen3.8-27B from @Alibaba_Qwen is open source, and Day-0 support is live in SGLang:
- 206.1 tok/s decode on a single RTX 5090, with our NVFP4 plus DSpark
- 38.28 tok/s decode on DGX Spark
Qwen3.8-27B raises the bar again for what a small model can do on agentic planning and long-horizon tasks.
Long live the (small model) king! Run it locally with SGLang 👇
Leandro Barbosa retweeted
Cursor is now part of @SpaceX.
Today, we have officially closed our acquisition. We will join the @spacexai team to help make Grok the world's most useful AI and improve Grok Build, Grok Bot, Grok API, Cursor, and more.
SpaceX has built some of the most inspiring and impressive technology in the world, and we’re grateful for the opportunity to become part of such a special company. Onwards.
1.58-bit text to speech. Weights of only -1, 0, +1. It speaks Brazilian Portuguese and English.
One video, English then Portuguese. Each plays full precision first, then the 1.58-bit version, same sentence and voice. The label switches at the seam. Can you hear a difference?
1
1
46
The unlock was mixed precision: a tiny precision-critical slice stays FP16, the rest goes ternary. Which slice, for speech, is the hard-won part. Wrong slice and it is just noise.
1
25
Portuguese WER around 0.11 to 0.13, matching full precision. English around 0.92. It runs on your own machine and nothing leaves the box. Built on F5-TTS by the X-LANCE lab at SJTU (Chen et al., 2024). github.com/SWivid/F5-TTS
@alexocheema @exolabs @huggingface
24
1.58-bit text-to-video.
We ternarized Wan2.2-TI2V-5B's FFN layers to weights of just -1/0/+1 and ran it at native 720p.
Left: FP16. Right: 1.58-bit.
Can you tell which is which?
#genai #quantization #diffusionmodels #BitNet
1
161
Leandro Barbosa retweeted
Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API.
Our ‘Fugu Ultra’ model matches the performance of Fable and Mythos, delivering frontier capability without the risk of export controls.
Try it: sakana.ai/fugu 🐡
Leandro Barbosa retweeted
16 parallel runs of Gemma 4 26B A4B on a single NVIDIA DGX Spark!
Pushing 18 tok/s per instance and a 300 tok/s aggregate. It can even hit 32 parallel runs.
This level of concurrency highlights how efficient the architecture is.
Leandro Barbosa retweeted
1-bit GLM-5.2 GGUF vs. Claude 4.8 Opus vs. GPT-5.5
We gave 3 models the same prompt and compared one-shot outputs.
The 1-bit GLM-5.2 GGUF ran locally on a Mac Studio M3 Ultra with 256GB RAM at ~21.6 tok/s.
Which output do you like best?
GGUF: huggingface.co/unsloth/GLM-5…
GLM-5.2 can now be run locally!🔥
The 2-bit model retains ~82% accuracy after we shrunk it from 1.51TB to 238GB (-84% size).
Run on a 256GB Mac or RAM/VRAM setups.
GLM-5.2 is the strongest open model to date.
Guide: unsloth.ai/docs/models/glm-5…
GGUF: huggingface.co/unsloth/GLM-5…





















