Elon Musk, scratched by K-startup AI achievements… “Still, we are number one.”

"Still, Grok is No. 1"… Elon Musk Contrasts with K-startup AI Model

Elon Musk, CEO of Tesla, has stirred interest by posting a contrasting comment after a South Korean startup's artificial intelligence (AI) large language model (LLM) received high praise.

On the 19th (local time), Elon Musk shared a post on his platform X (formerly Twitter) referencing a post by the American AI performance analysis agency 'Artificial Analysis' from the previous day, stating, "But xAI's Grok is still number one. And Grok is improving rapidly."

At this time, Artificial Analysis is used as a key reference point by experts in the AI industry.

Image 1

The post Musk shared indicated that the new reasoning AI model Solar Pro 2 from the South Korean startup Upstage was evaluated as the 12th best performing AI model globally, noting that it "shows impressive performance for its size and has very high cost competitiveness."

The new reasoning AI model Solar Pro 2 from South Korean startup Upstage

"It shows impressive performance for its size and has very high cost competitiveness."

In a previous post, Artificial Analysis selected Solar Pro 2 as 'the most cost-competitive model,' while rating Grok 4 as 'the model with the lowest cost competitiveness.'

As a result, Musk made a comment seemingly acknowledging Solar Pro 2 for its notable achievements in performance and cost-effectiveness.

Kim Seong-hoon, CEO of Upstage, commented on the post, "Congratulations to xAI for being number one. But it won't last long. Upstage's Solar LLM will improve rapidly and catch up," expressing confidence.

"Congratulations to xAI for being number one. But it won't last long."

Image 2

Upstage's Solar Pro 2, launched on the 10th, is designed with 31 billion parameters, making it an optimal performance reasoning model.

In particular, it is noted for being the first Upstage LLM to apply a 'hybrid mode,' allowing users to choose between 'chat mode,' optimized for quick responses, and 'reasoning mode,' which provides structured answers through step-by-step thinking.

The first Upstage LLM to apply 'hybrid mode.'

Artificial Analysis also ranked the developer, Upstage, among the 'Top 10 Companies Creating Frontier Language Models' alongside OpenAI, xAI, Meta, and Google.

Image 3

Image Source: Referenced material for understanding the article / gettyimageskorea, x 'hunkims', Solar Pro 2 / Captured from Upstage's website