https://tpsreport.news/news/ibm-granite-4-2-8b-dense-reasoning-model TPS — Tokens Per Second en 2026-08-31T20:35:43.546Z IBM Releases Granite 4.2 8B, a Dense Reasoning Model with 131K Context and Three Thinking Modes IBM, Granite, Granite 4.2, dense model, reasoning model, open weights, multilingual https://tpsreport.news/news/aws-multi-tenant-agentic-chat-bedrock-knowledge-base TPS — Tokens Per Second en 2026-08-31T19:05:42.977Z AWS Details Reference Architecture for Multi-Tenant Document Chat on Amazon Bedrock Knowledge Bases Amazon Bedrock, AWS, Knowledge Base, RAG, multi-tenant, agentic retrieval, enterprise AI https://tpsreport.news/news/perplexity-personal-computer-agent-windows TPS — Tokens Per Second en 2026-08-31T17:05:40.440Z Perplexity Brings Agentic 'Personal Computer' Feature to Windows Perplexity, agentic AI, Windows, product update, AI agents, automation https://tpsreport.news/news/openai-outcome-based-pricing-enterprise TPS — Tokens Per Second en 2026-08-31T14:35:45.472Z OpenAI Quietly Rolls Out Outcome-Based Pricing, Charging Some Customers Only When Tasks Succeed OpenAI, pricing, enterprise AI, outcome-based pricing, Salesforce, Anthropic, business model https://tpsreport.news/news/deepseek-v4-flash-vision-exp-multimodal TPS — Tokens Per Second en 2026-08-31T11:51:22.517Z DeepSeek Releases V4-Flash-Vision-Exp, First Multimodal Model in V4 Family DeepSeek, DeepSeek-V4, multimodal, open-source, vision-language-model, Hugging Face https://tpsreport.news/news/ai-agents-no-sense-of-time-study TPS — Tokens Per Second en 2026-08-30T10:50:51.717Z Study Finds AI Coding Agents Cannot Track Elapsed Time or Judge Their Own Work Quality AI agents, coding assistants, Claude Code, OpenAI Codex, AI research, agent evaluation, MATS https://tpsreport.news/news/claude-code-usage-limit-change-net-cut TPS — Tokens Per Second en 2026-08-30T09:20:39.726Z Anthropic to Cut Claude Code Weekly Limits by 17% Despite Calling It a 25% Increase Claude Code, Anthropic, usage limits, AI coding assistant, pricing