TechCrunch2 min read

By Interestana AI EditorialAI summary of reporting by TechCrunchHow we report

Moonshot's Kimi 3 Aims to Match Anthropic's Opus

Moonshot's forthcoming Kimi 3 model is anticipated to significantly narrow the performance gap with Anthropic's Claude 3 Opus, according to a report by the Financial Times. Kimi 3 is slated to be China's largest open artificial intelligence model, boasting a parameter count estimated to be between 2 trillion and 3 trillion. This substantial scale suggests a considerable advancement in the capabilities of Chinese AI development.

The development of Kimi 3 positions it as a potential competitor to leading global AI models. Anthropic's Claude 3 Opus, released earlier in 2024, has set a high benchmark for multimodal reasoning and complex task completion. By aiming to match Opus 4.8, Moonshot signals its ambition to compete at the forefront of AI research and deployment.

The Financial Times report highlights that Kimi 3's parameter count places it among the largest models globally. The specific figure of 2 to 3 trillion parameters indicates a significant increase in model complexity and potential for sophisticated understanding and generation. This scale is crucial for advanced AI tasks, including nuanced language comprehension, complex problem-solving, and creative content generation.

Moonshot's strategy with Kimi 3 appears focused on leveraging open-source principles to accelerate AI development and accessibility within China and potentially globally. The success of Kimi 3 could mark a pivotal moment for China's AI industry, demonstrating its capacity to produce models that rival those from established international players like Anthropic.

Original source — read the full reporting at the publisher:

Read on TechCrunch(opens in a new tab)

Ref 005PTTYR

Follow this story

Get Artificial Intelligence in your inbox — each new story in 40 words or fewer, daily or weekly.

Subscribe

Read next