Kimi K3, and what we can still learn from the pelican benchmark

·Simon Willison··

Chinese AI lab Moonshot AI announced Kimi K3 this morning, describing it as their "most capable model to date, with 2.8 trillion parameters". It's currently available via their website and API, but an open weight release is promised "by July 27, 2026". Moonshot are calling this the first "open 3T-class model" (I guess they're rounding 2.8 trillion up to 3 trillion), taking the crown from DeepSeek's 1.6T v4 Pro. Their self-reported benchmarks have K3 mostly beating Claude Opus 4.8 max and GPT-5.5...

Read full article →

Related Articles

Should I run plain Docker Compose in production in 2026?
pmig · Hacker News · 2mo ago
Computer Use is 45x more expensive than structured APIs
palashawas · Hacker News · 2mo ago
Bun is being ported from Zig to Rust
SergeAx · Hacker News · 2mo ago
Show HN: Tilde.run – Agent sandbox with a transactional, versioned filesystem
ozkatz · Hacker News · 2mo ago
RaTeX: KaTeX-compatible LaTeX rendering engine in pure Rust
atilimcetin · Hacker News · 2mo ago