_saurabh's profile picture. Leading code intelligence @ Essential AI

Previously: 2x YC (W15, S18); PhD + Postdoc in Program Synthesis

Saurabh Srivastava

@_saurabh

Leading code intelligence @ Essential AI Previously: 2x YC (W15, S18); PhD + Postdoc in Program Synthesis

Disematkan

Code is an amazing petri dish for measuring and building intelligence. We released Rnj-1 yesterday; a 8B code & science model that you can run offline on your laptop. It got 20.8% (!) on the SWEBench software engineering benchmark. For context: a) that beats all comparable 8B…

_saurabh's tweet image. Code is an amazing petri dish for measuring and building intelligence.

We released Rnj-1 yesterday; a 8B code & science model that you can run offline on your laptop. It got 20.8% (!) on the SWEBench software engineering benchmark. 

For context: a) that beats all comparable 8B…

Saurabh Srivastava memposting ulang

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

ashVaswani's tweet image. Rnj-1-Instruct is now the #1 trending text generation model on HF!

445k downloads in two days. appreciate that people are trying Rnj-1. In top-10 amongst all models and #1 in small text-gen models.

_saurabh's tweet image. 445k downloads in two days. appreciate that people are trying Rnj-1. In top-10 amongst all models and #1 in small text-gen models.

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

essential_ai's tweet image. We are now the #1 trending text-gen &amp;lt;256B size model on HuggingFace!!


Saurabh Srivastava memposting ulang

NEW: @essential_ai just released Rnj-1, their first 8B model. You can easily fine-tune it with GRPO using TRL to add reasoning capabilities to a compact mode Free Colab link below

SergioPaniego's tweet image. NEW: @essential_ai just released Rnj-1, their first 8B model. 

You can easily fine-tune it with GRPO using TRL to add reasoning capabilities to a compact mode

Free Colab link below

Saurabh Srivastava memposting ulang

.@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! 🚀🚀🚀

ollama's tweet image. .@essential_ai&apos;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&apos;s go! 🚀🚀🚀

Saurabh Srivastava memposting ulang

Rnj-1 is now live on @OpenRouterAI!

essential_ai's tweet image. Rnj-1 is now live on @OpenRouterAI!

Saurabh Srivastava memposting ulang

sweet we're trending on @huggingface ! ^^

gordic_aleksa's tweet image. sweet we&apos;re trending on @huggingface ! ^^

Saurabh Srivastava memposting ulang

seems like the open stack is glm-4.6V-Flash for multimodal olmo 3 32B thinking for math rnj-1-instruct for agentic


Saurabh Srivastava memposting ulang

Congratulations to @ashVaswani and the @essential_ai team for advancing American open source AI! I hope this serves as a template for more startups to follow.

Today, we’re excited to introduce Rnj-1, @essential_ai's first open model; a world-class 8B base + instruct pair, built with scientific rigor, intentional design, and a belief that the advancement and equitable distribution of AI depend on building in the open. We bring…

essential_ai's tweet image. Today, we’re excited to introduce Rnj-1, @essential_ai&apos;s  first open model; a world-class 8B base + instruct pair, built with scientific rigor, intentional design, and a belief that the advancement and equitable distribution of AI depend on building in the open.

We bring…


Saurabh Srivastava memposting ulang

We are beyond thrilled to share our first flagship models, Rnj-1 base and instruct 8B parameter models. Rnj-1 is the culmination of 10 months of hard work by a phenomenal team, dedicated to advancing American SOTA OSS AI. Lots of wins with Rnj-1. 1. SWE bench performance close…

Today, we’re excited to introduce Rnj-1, @essential_ai's first open model; a world-class 8B base + instruct pair, built with scientific rigor, intentional design, and a belief that the advancement and equitable distribution of AI depend on building in the open. We bring…

essential_ai's tweet image. Today, we’re excited to introduce Rnj-1, @essential_ai&apos;s  first open model; a world-class 8B base + instruct pair, built with scientific rigor, intentional design, and a belief that the advancement and equitable distribution of AI depend on building in the open.

We bring…


Saurabh Srivastava memposting ulang

Today, we’re excited to introduce Rnj-1, @essential_ai's first open model; a world-class 8B base + instruct pair, built with scientific rigor, intentional design, and a belief that the advancement and equitable distribution of AI depend on building in the open. We bring…

essential_ai's tweet image. Today, we’re excited to introduce Rnj-1, @essential_ai&apos;s  first open model; a world-class 8B base + instruct pair, built with scientific rigor, intentional design, and a belief that the advancement and equitable distribution of AI depend on building in the open.

We bring…

Saurabh Srivastava memposting ulang

Just converted Rnj-1 to mlx versions 4Bit, 8Bit and fp16 in case anyone is interested in testing. Rnj-1 looks promising for an open model. It's only an 8B parameter base model so small and mighty to start with. @huggingface really makes open source equitable and attainable to…

alexcovo_eth's tweet image. Just converted Rnj-1 to mlx versions 4Bit, 8Bit and fp16 in case anyone is interested in testing. Rnj-1 looks promising for an open model. It&apos;s only an 8B parameter base model so small and mighty to start with.

@huggingface really makes open source equitable and attainable to…

Today, we’re excited to introduce Rnj-1, @essential_ai's first open model; a world-class 8B base + instruct pair, built with scientific rigor, intentional design, and a belief that the advancement and equitable distribution of AI depend on building in the open. We bring…

essential_ai's tweet image. Today, we’re excited to introduce Rnj-1, @essential_ai&apos;s  first open model; a world-class 8B base + instruct pair, built with scientific rigor, intentional design, and a belief that the advancement and equitable distribution of AI depend on building in the open.

We bring…


Saurabh Srivastava memposting ulang

range-1 is a pretty sick small model for agent/function calling use-cases! 🔥 >> Given this is from Vaswani & team, iam looking forward to seeing if theres any spicy Transformer modifications behind this model 👀🌶️🌶️🌶️ >> First OSS 🇺🇸 model competitive with Qwen3-8b >> fully…

We are beyond thrilled to share our first flagship models, Rnj-1 base and instruct 8B parameter models. Rnj-1 is the culmination of 10 months of hard work by a phenomenal team, dedicated to advancing American SOTA OSS AI. Lots of wins with Rnj-1. 1. SWE bench performance close…



Loading...

Something went wrong.


Something went wrong.