DeepSeek's Flash update claims gains on 11 benchmarks at the same price
DeepSeek's V4.1-Flash-1017 reports higher scores on all 11 benchmarks, with open MIT weights and unchanged prices. TECHi checks the costs and memory math.
Tag
#AI Models
7 articles
TECHi reporting and analysis covering AI Models. 7 articles, newest first.
Anthropic raises misalignment risk label after cyber disclosures
Anthropic’s August report moved misalignment risk from Very Low to Low, citing greater uncertainty after cyber disclosures.
Fenghe Is Open Source. Its Forecasting Proof Is Not.
CMA's 106B-parameter Fenghe model is open for inspection, but its release does not independently verify forecasting skill or operational safety at scale.
Kimi K3 Is Cheap to Call, but the Weights Are Not Here
Kimi K3’s API is live, but its weights, license and self-hosting stack are not. TECHi audits the 64-accelerator gap behind Moonshot AI’s claims today.
Inkling’s Open Weights Still Need a Data Center
Thinking Machines Lab's Inkling publishes open weights, but its 600 GB VRAM floor, limited fine-tuning contexts and launch-day gaps complicate ownership.
AI progress so far, and the next 10 years that matter
A research-based forecast of AI progress through 2036, covering agents, science, jobs, energy, regulation, and infrastructure.
Tencent’s Hy3 preview Marks the First Major Model to Emerge From Its AI Rebuild
Tencent has introduced Hy3 preview, the first major model to emerge since the company rebuilt key parts of its pre-training and reinforcement-learning infrastructure earlier this year, making the release as much a sign of internal change as a routine model update. According to briefing materials, Hy3 preview reached pu