Ali Shehper

RS @ NVIDIA | PhD in Theoretical Physics

[Andreas Karch @karch_andreas]

Proving BFSS has a unique groundstate is really something (especially to us here at UT, home of F). But what are these extra finite energy boundstates doing at N=2?

[Zohar Komargodski @ZoharKo]

Many great results from OpenAI. Congratulations!
This one in particular is very close to my heart. A decade or so ago, with collaborators, I have published a physics-oriented proof of it.

https://t.co/fzOQHGe8z2

[NVIDIA AI @NVIDIAAI]

Congratulations to the students who competed at the International Mathematical Olympiad (IMO) 2026. 👏

We put Nemotron 3 Ultra to the test to take on the same problems in the same time limit, with no internet or external tools.

The IMO team graded its solutions 30/42, above the 29-point gold threshold. 🥇

[NVIDIA AI @NVIDIAAI]

Today we're shipping Nemotron 3 Ultra.

A 550B MoE frontier-intelligence open model built for long-running agents.

It delivers 5x faster inference and lowers the cost of complex agentic tasks by up to 30% versus other open frontier models.

[Nathan Lambert @natolambert]

Nvidia is the tip of the spear for the growing US open model efforts.

Is a long time coming, but obvious when a 550B model drops that more people will take notice.

Many people won’t realize the valuable training data they release too.

[Josh @JMRLudan]

this is what reading RL papers feels like

[Ashish Vaswani @ashVaswani]

Rnj-1 aces telco benchmarks. Congratulations to Farbod Tavakkoli, @GregoryDiamos , @tensorwave , @AMD , and the @essential_ai team!

[Essential AI @essential_ai]

Rnj-1 has outperformed other open models in its weight class in the largest open-source AI initiative in telecom to-date🚀

[Essential AI @essential_ai]

[1/5]

We thank the community for their feedback on Rnj-1. We’d like to announce a few updates to Rnj-1-instruct based on what we heard:

- Resolving premature truncation of generations and improved instruction following.
- Instructions for 128k context length extrapolation.
- Updated evals, baselines, and model generations for reproducibility.

Details follow 🧵

[Ashish Vaswani @ashVaswani]

Rnj-1-Instruct is now the #1 trending text generation model on HF!

[Essential AI @essential_ai]

The momentum is real. Let's go! 🚀

[Essential AI @essential_ai]

We are now the #1 trending text-gen <256B size model on HuggingFace!!

[ollama @ollama]

.@essential_ai's rnj-1 model is now on Ollama!

ollama run rnj-1

8B parameter, open-weight dense model trained from scratch. The model is optimized for code and STEM with capabilities on par with other state of the art open-weight models.

Let's go! 🚀🚀🚀

[Aleksa Gordić (水平问题) @gordic_aleksa]

sweet we're trending on @huggingface ! ^^

[Essential AI @essential_ai]

Rnj-1 is now live on @openrouter!

[Ameya Velingker | अमेय वेलिंगकर @ameya_pa]

Benchmark performance tends to get the bulk of attention with AI model releases, but there a few cool underdiscussed gems about Rnj-1:

- We trained the 8B models directly *from scratch*, not by training and distilling a larger model to a smaller one.
- Post-training for Rnj-1 Instruct was light, i.e., it involved simple SFT but no RL, DPO, etc.
- Our infra team worked out a lot of infra challenges to make our unified JAX training framework work seamlessly across disparate hardware platforms: both AMD GPUs and TPUs.

[Ali Shehper @ashehper]

@gordic_aleksa @ameya_pa @kurtosis0 @mobydobius @Khoi_H_Nguyen 😂

[Ameya Velingker | अमेय वेलिंगकर @ameya_pa]

Indeed, it has been a pleasure getting the chance to work with such fantastic colleagues in the STEM team (including the two people who weren't in town for the photo).

@kurtosis0 @mobydobius @gordic_aleksa @ashehper @Khoi_H_Nguyen

[Ashish Vaswani @ashVaswani]

It was a wonderful experience working with @vipulved and the @togethercompute team to bring Rnj-1 to the world.

[Yupp @yupp_ai]

📢 New Model Drop: Rnj-1 Instruct is now live on Yupp!

The first flagship model from @essential_ai, this open-source model is built to excel at math problem-solving and scientific reasoning.

We checked it out with some prompts: