open-weight model
4 published articles
Open-weight reasoning
Nvidia's Nemotron-4 runs 30% cheaper than GPT-4. The catch is buried in the benchmark.
Nvidia's Nemotron-4 line challenges Llama 3.3 and Mistral Large with competitive MMLU scores and a claimed 30% inference savings. The open license and dual-size strategy position it as a practical alternative for teams running agentic workloads at scale, pending third-party verification.
2026-07-25
Artificial Intelligence
The $314 billion assumption that just broke
Moonshot AI's Kimi K3, a 2.8 trillion parameter open-weight model, triggered a global revaluation of AI stocks. The combined estimated valuations of OpenAI and Anthropic lost $314 billion as investors confronted cheaper Chinese frontier AI. The analysis examines winners (Chinese chip makers, memory manufacturers, software developers) and losers (model developers, Nvidia's scarcity premium) in the new landscape.
2026-07-22
Artificial Intelligence
MiniMax's M3 just beat Opus 4.7 at browsing, trained itself, and never asked for help
MiniMax M3 delivers a 9.4x CUDA kernel speedup, beats Opus 4.7 on BrowseComp, and autonomously replicated an ICLR paper. All in an open-weight package, and it never asked for help.
2026-07-06
Model Release
DeepSeek-V4 preview lands, and the open-weight math gets harder for everyone
DeepSeek unveils DeepSeek-V4 preview with top-tier reasoning and stronger agent abilities, now available across platforms. The release signals a major escalation in the open-weight AI race, putting pressure on OpenAI and Google to justify their closed-source strategies.
2026-07-05