The records behind this event
2-
AI Model
Grok 4
xAI's frontier reasoner: Humanity's Last Exam leader at launch.
Grok 4 (July 2025) led hard-reasoning benchmarks including Humanity's Last Exam at release, with a Heavy multi-agent tier; Grok 4.1 (November 2025) tuned it for everyday use atop the X platform. Its rollout also featured high-profile alignment stumbles on X: a case study in deploying frontier models into a social feed. Details undisclosed. Point releases…
Open record → -
AI Model
Kimi K2
A trillion-parameter open MoE tuned for agentic work.
Moonshot's K2 (July 2025) is a 1T-parameter MoE (32B active) released with open weights and explicitly optimized for tool use and agentic coding; K2 Thinking (November 2025) added a reasoning tier that challenged closed frontier models on several benchmarks. Trained under export-control constraints, it is among the largest open models ever released. K2.6…
Open record →