Technology ❯ Artificial Intelligence

AI Benchmarks

Performance Metrics Humanity's Last Exam Nvidia Transparency General AI Assistants TRUEBench ARC-AGI-2 Job Automation DeepSearchQA SOTA Performance

Moonshot Launches Kimi K3.1 Pro With One‑Million‑Token Context and Highspeed Variant

The company says it reduced the model’s offensive cyber capability by removing cyber training data to balance wider access and safety.

Anthropic’s Claude Sonnet 5.5 Comes to Snowflake Cortex AI and Climbs Agent Leaderboard

Google Launches Android Bench to Rank AI Models on Real Android Coding Tasks

OpenAI Releases GPT-5.4 With Native Desktop Control and Unified Frontier Capabilities

Google Rolls Out Gemini 3.1 Flash‑Lite in Preview for High‑Volume AI Workloads

Google Launches Interactions API and Upgraded Gemini Deep Research for Developers

xAI Rolls Out Grok 4.1 Free to All Users, Citing Faster, More Accurate Replies

Moonshot AI Releases Open-Source Kimi K2 Thinking, a Tool-Using 'Thinking Agent'

OpenAI Releases GDPval Work Benchmark as Claude Leads Early Results

Samsung Unveils TRUEBench to Measure LLM Productivity in Real-World Tasks

Google Expands Gemini 2.5 Pro Experimental AI Access to Free Users

OpenAI Unveils 'Deep Research' AI Agent for Autonomous Web Investigations

Nvidia Dominates AI Benchmarks with Record-Breaking Performance