Moonshot's Kimi K3 set to challenge Anthropic's Opus 4.8
Chinese AI lab Moonshot AI is set to release Kimi K3 in the coming days, a model sources say could match or beat Anthropic's Opus 4.8. The Financial Times reports Kimi will be China's largest open-weight AI model, with 2 trillion to 3 trillion parameters, as Moonshot also seeks a higher valuation.
Key Takeaways
- Kimi K3 is expected within days and may match or surpass Anthropic's Opus 4.8, according to people cited by the Financial Times.
- The model is positioned as China's largest open-weight release, with a reported 2–3 trillion parameters.
- Moonshot is reportedly raising capital at a $31.5 billion valuation after a May round that valued it at $20 billion.
- A July 15 teaser and Arena.ai "Kivine" sightings have fueled comparisons with Claude Fable5.
What is Moonshot launching with Kimi K3?
According to TechCrunch, citing the Financial Times, Moonshot AI's next Kimi model—Kimi K3—is due "in the coming days."
FT sources say it will be the largest open-weight AI model from China, sized between 2 trillion and 3 trillion parameters. Earlier Kimi K2 models already ranked high on open-source benchmarks and trailed frontier systems by a smaller margin than many rivals.
The release is framed as closing the gap with closed-source leaders from Anthropic and OpenAI. For more viral tech and culture coverage, see Celebrity Breaking News on BlasterPost.
How does Kimi stack up against Anthropic's models?
Anonymous sources told the FT that Kimi K3 is expected to perform on par with—or even surpass—Anthropic's Opus 4.8. That claim has not been independently verified ahead of launch.
Separately, AIBase reported that Moonshot posted a July 15 teaser across Bilibili, X, and Xiaohongshu. An anonymous Arena.ai model codenamed Kivine is widely speculated to be tied to Kimi K3, and netizens have shared side-by-side tests against Claude Fable5 and GPT-5.6 Sol.
Moonshot's prior K2 line, including April 2026's open-sourced K2.6 with support for up to 300 parallel sub-agents, built momentum for this challenge in coding and agent workloads.
Why does the Kimi launch matter for AI buyers?
The timing feeds a wider debate over paying premium prices for closed models such as Claude and ChatGPT. Industry leaders worry about how client data is used, and some executives urge firms to adopt cheaper open-weight options from labs including DeepSeek, Z.ai, or Moonshot and customize them in-house.
Moonshot is also said to be raising fresh capital at a $31.5 billion valuation, after raising $2 billion in May at a $20 billion valuation. If Kimi K3 delivers on the reported ambitions, open-weight models from China could put fresh pressure on U.S. frontier labs' pricing and positioning.