#Open Weights

Kimi K3 Runs on 20,000 Rented Nvidia Chips

Moonshot's Alibaba compute deal shows how Chinese labs actually get frontier silicon, and why Kimi builders should care.

Article · 1w ago2

Kimi K3 Turns a Year of Efficiency Tricks Into 2.8T Parameters

Moonshot's new flagship validates hybrid attention, latent MoE, and 4-bit training at frontier scale — and redefines open weights.

Article · 2w ago1

Anthropic's Open-Weights Position Is Narrower Than It Looks

Amodei's three policy asks target frontier releases and industrial distillation, not the weights already on your GPU.

Article · 2w ago5

Kimi K3 Is Open. You Still Can't Run It.

Moonshot's 2.8-trillion-parameter release is the biggest open-weights drop ever, and it redefines what open means.

Article · 2w ago2

Big Tech's Open-Weights Letter Is Really About Chinese Models

Nvidia, Microsoft, Meta and 22 others move to block a ban that would strand the startups built on Chinese weights.

Article · 2w ago1

Hetzner Is Quietly Commoditizing LLM Inference

A free, OpenAI-compatible endpoint from Europe's cheapest host says more about inference margins than any benchmark.

Article · 3w ago1

Cisco's $1 Bug Finder Ties GPT-5.5 by Undercutting It

Antares shows tiny specialized models matching frontier ones at vulnerability triage — and reveals how unsolved the task still is.

Article · 3w ago7

Kimi K3 Catches Fable, but the 50x Cost Claim Doesn't Add Up

The open-weights gap just collapsed to one generation; the real frontier moat is now token efficiency.

Article · 3w ago5

The Week Open Chinese Models Called Washington's Bluff

A record open-weight release, Xi's open-source pitch, and a breach where the only usable model was Chinese.

Article · 3w ago0

Kimi K3 Is the Bill for America's Closed-AI Bet

Chinese labs now own the open-weights layer American startups build on, and last week made that expensive to ignore.

Article · 3w ago1

Inkling Puts 975B Open Weights in Reach

Murati's lab ships a frontier MoE you can fine-tune, not just rent.

Article · 4w ago2

Old Xeons Can Run Gemma 4 at Reading Speed

MoE math, CPU kernels, and speculative decoding turn retired enterprise boxes into real local inference hosts.

Article · 4w ago1

The Open-Weights Gap Depends on What You Measure

A viral chart predicts open models reach parity by December 2026. Across 18 benchmarks, the honest answer is messier.

Article · 1mo ago5

The Real Cost of the Open-Weight Price Collapse

Why the 50x price gap between open-weight APIs and frontier models is rewriting the build-versus-buy playbook for developers.

Article · 1mo ago2

GLM 5.2 Is a Point Behind Opus — Until the Task Runs for Hours

Open weights and a 5.7x price cut make the near-parity headline real, but the gap reopens exactly where long-horizon agents live.

Article · 1mo ago2