Kimi-K3 narrows the open-source gap to closed frontier models to about 1.5 months
ImaginaryRea1ity · reddit · 2026-07-21
- The post argues that Kimi-K3 is still not better than Fable, but is clearly closing the gap.
- It says the open-source frontier is now only about 1.5 months behind closed models, based on the Artificial Analysis chart shown in the image.
- The chart compares the evolution of closed vs open-weights frontier models and highlights the rapid growth of recent open models.
- The author also points out that the biggest open models are getting larger, with Kimi-K3 reportedly nearing 3T parameters, so it is far from something that runs comfortably on a laptop.
- The image also raises a question about Google’s relative pace since Gemini 3 Pro last November.
Related event: Moonshot Releases 2.8T Open-Weights Model Kimi K3(14 posts)→
More from Models
- ChatGPT keeps telling people 404 Media is tech support for big corporations — 404 Media · 2026-10-26
- Claude's /show-me visualizes compressor auto makeup gain like a pro audio teacher — teropa · 2026-10-26
- ChatGPT keeps telling people 404 Media can fix their tech support problems — 404 Media · 2026-10-26
- Fern's "context is intelligence" thesis: prediction keeps improving out to 10.2M tokens — bneyshabur · 2026-10-26
- Mirendil launches Fern, a from-scratch pretrained model with up to 10M-token context — bneyshabur · 2026-10-26
- Users spot a dedicated translation UI appearing in Claude — BLCNYY · 2026-10-26