7 articles tagged
Mistral released Shieldstral 1.0 as Apache 2.0 open weights on 4 August 2026. On Mistral's own benchmark tables, GPT-OSS-Safeguard-20B leads 9 of 18 text rows to Shieldstral's 4.
Open weights closed the intelligence gap from 13 points to 6, yet enterprise share fell to 11%. New August 2026 data shows identical weights scoring 22% or 37% on tool calling depending only on the endpoint.
Qwen3.8-Max shipped on 3 August 2026 at $2 input and $6 output per million tokens, with a 1M context and 15,000 requests per minute. But Qwen3.7-Max is on a 50% discount at $1.25 and $3.75, so the newer flagship costs
On 27 July 2026 NVIDIA and 36 partners launched the Open Secure AI Alliance and open-sourced NOOA, days after closed APIs blocked forensics in the Hugging Face breach. A guide to open vs closed AI-agent security.
Kimi K3 (2.8T), DeepSeek V4 (1.6T) and GLM-5.2 (753B) are all open-weight and frontier-class. This decision matrix compares GPU footprint, API list price, and the monthly output-token volume where self-hosting each one
Thinking Machines released Inkling on 15 July 2026 under Apache 2.0: 975B parameters, 41B active. We priced the GPUs against the Tinker API using AWS and vendor figures, and the break-even is brutal.
Chinese-built AI models now run 30 to 46% of US enterprise API tokens, up from about 4.5% in early 2025, driven by prices 60 to 90% below OpenAI and Anthropic. The cost case and the governance risks to weigh.
