https://tpsreport.news/news/nvidia-harness-research-claude-opus-5-arc-agi-3 TPS — Tokens Per Second en 2026-08-21T19:51:23.108Z Nvidia Research: Custom Harness Pushes Claude Opus 5 from 30% to 100% on ARC-AGI-3 nvidia, agentic-ai, ARC-AGI-3, Claude Opus 5, AI agents, benchmarks, harness https://tpsreport.news/news/anthropic-claude-security-mythos-5-cyber-defense TPS — Tokens Per Second en 2026-08-21T19:50:56.356Z Anthropic Deploys Claude Mythos 5 in New Security Scanner for Enterprise Codebases Anthropic, Claude Mythos 5, cybersecurity, enterprise AI, Claude Security https://tpsreport.news/news/deepseek-v4-flash-vision-exp-multimodal-agent-model TPS — Tokens Per Second en 2026-08-21T19:20:46.681Z DeepSeek Releases Experimental V4-Flash-Vision-Exp, Claims Near-Parity With Opus 4.8 on Agent Benchmarks DeepSeek, V4-Flash, multimodal, vision model, AI agents, benchmarks, Opus 4.8 https://tpsreport.news/news/grok-bot-xai-cursor-agentic-ai-app TPS — Tokens Per Second en 2026-08-21T18:36:07.621Z xAI and Cursor Launch Grok Bot, an Agentic AI App for Mac and iPhone Grok, xAI, Cursor, AI agent, iOS app, Mac app, agentic AI https://tpsreport.news/news/chatgpt-mac-imessage-plugin TPS — Tokens Per Second en 2026-08-21T18:05:45.722Z OpenAI Launches ChatGPT Plugin That Reads and Analyzes Mac iMessages ChatGPT, OpenAI, Mac, iMessage, plugin, privacy, Apple https://tpsreport.news/news/github-copilot-slack-integration-public-preview TPS — Tokens Per Second en 2026-08-21T16:20:56.097Z GitHub Brings Agentic Copilot CLI Capabilities to Slack in Public Preview github-copilot, slack, product-update, developer-tools, agentic-ai, public-preview https://tpsreport.news/news/github-copilot-shared-agentic-work-microsoft-teams TPS — Tokens Per Second en 2026-08-21T16:20:38.001Z GitHub Copilot Now Works as a Shared Agent Inside Microsoft Teams Channels GitHub Copilot, Microsoft Teams, product update, agentic AI, developer tools https://tpsreport.news/news/deepseek-v4-flash-vision-exp-multimodal-moe TPS — Tokens Per Second en 2026-08-21T12:06:34.100Z DeepSeek Releases V4 Flash Vision Exp, an Experimental Multimodal MoE Model with 1M Context DeepSeek, DeepSeek V4, multimodal, vision model, mixture-of-experts, open model, OpenRouter https://tpsreport.news/news/deepmind-eve-online-game-ai-research-sima-2 TPS — Tokens Per Second en 2026-08-21T12:06:07.582Z Google DeepMind Extends Game AI Research to EVE Online, Building on SIMA 2 Agent Google DeepMind, SIMA 2, reinforcement learning, game AI, EVE Online, AlphaGo, research partnership https://tpsreport.news/news/chatgpt-imessage-plugin-mac TPS — Tokens Per Second en 2026-08-21T10:20:42.371Z OpenAI Launches ChatGPT Plugin That Reads and Replies to iMessages on Mac OpenAI, ChatGPT, Apple, iMessage, macOS, privacy, product integration https://tpsreport.news/news/anthropic-data-retention-policy-change-enterprise TPS — Tokens Per Second en 2026-08-21T10:05:42.307Z Anthropic Reverses Course, Will Let Enterprise Customers Store Retention Data on Their Own Cloud Anthropic, data retention, enterprise AI, data privacy, cloud storage, compliance https://tpsreport.news/news/gpt-image-2-transparent-background-api TPS — Tokens Per Second en 2026-08-21T08:05:38.671Z OpenAI Adds Transparent Background Generation to GPT-Image-2 API OpenAI, GPT-Image-2, image generation, API update, transparent background https://tpsreport.news/news/cline-v4-1-11-release TPS — Tokens Per Second en 2026-08-21T05:50:40.260Z Cline v4.1.11 Adds Inline Image Generation, Fixes VS Code 1.134 Compatibility Break Cline, VS Code, coding agent, changelog, developer tools https://tpsreport.news/news/chatgpt-apple-messages-plugin TPS — Tokens Per Second en 2026-08-20T22:20:36.389Z OpenAI Launches Apple Messages Plugin, Letting ChatGPT Read, Draft, and Send Texts OpenAI, ChatGPT, Apple, Messages, plugin, privacy, Codex, ChatGPT Work https://tpsreport.news/news/gpt-5-6-cross-region-inference-amazon-bedrock TPS — Tokens Per Second en 2026-08-20T22:05:42.893Z AWS Bedrock Adds Cross-Region Inference for OpenAI's GPT-5.6 Models Amazon Bedrock, OpenAI, GPT-5.6, AWS, cross-region inference, cloud infrastructure https://tpsreport.news/news/pixel-11-gboard-personalized-suggestions-gemini TPS — Tokens Per Second en 2026-08-20T21:21:11.050Z Pixel 11's Gboard Adds Gemini-Powered 'Personalized Suggestions' That Read Screen Context Gboard, Pixel 11, Gemini, Google, Private AI Compute, keyboard AI, Android https://tpsreport.news/news/ox-alpha-stealth-model-openrouter TPS — Tokens Per Second en 2026-08-20T21:20:46.488Z Anonymous 'Ox Alpha' Reasoning Model Appears on OpenRouter with Free 1M-Token Context stealth model, OpenRouter, reasoning model, Ox Alpha, agentic coding, long context https://tpsreport.news/news/chatgpt-apple-messages-integration-mac TPS — Tokens Per Second en 2026-08-20T18:05:47.609Z OpenAI Adds Apple Messages Plugin to ChatGPT for Mac ChatGPT, OpenAI, Apple Messages, macOS, product update, Codex, privacy https://tpsreport.news/news/lfm25-dspark-speculative-decoding-speedup TPS — Tokens Per Second en 2026-08-20T17:05:47.595Z Liquid AI Ships DSpark Draft Models, Cutting LFM2.5 Inference Latency Up to 3.18x on GPU Liquid AI, LFM2.5, speculative decoding, inference optimization, llama.cpp, SGLang, on-device AI, DSpark https://tpsreport.news/news/aws-bedrock-agentcore-dogwood-policy-authoring TPS — Tokens Per Second en 2026-08-20T16:50:41.525Z AWS Expands Natural-Language Policy Authoring for AI Agents in Amazon Bedrock AgentCore AWS, Amazon Bedrock, AgentCore, AI agents, governance, Dogwood, policy enforcement, guardrails https://tpsreport.news/news/slack-code-ai-agents-group-chat TPS — Tokens Per Second en 2026-08-20T15:35:44.763Z Slack Launches 'Code Channels' to Put AI Coding Agents Into Group Chats Slack, Salesforce, AI agents, coding agents, product update, Anthropic, GitHub Copilot, OpenAI https://tpsreport.news/news/mistral-agentic-search-retrieval-layer TPS — Tokens Per Second en 2026-08-20T14:21:22.888Z Mistral Launches Agentic Search, Claims 3x Accuracy Gain on Financial Document Retrieval Mistral, agentic search, RAG, retrieval, enterprise AI, FinanceBench https://tpsreport.news/news/china-ai-catch-up-frontier-radar-analysis TPS — Tokens Per Second en 2026-08-20T14:20:56.988Z Chinese Models Kimi K3 and GLM-5.3 Close In on GPT-5.5 and Claude Opus 5, New Analysis Finds China AI, Kimi K3, GLM-5.3, Qwen3.8-Max, Anthropic, OpenAI, benchmarks, AI safety https://tpsreport.news/news/tencent-hy-mt2-30b-a3b-translation-model TPS — Tokens Per Second en 2026-08-20T13:35:49.220Z Tencent Releases Hy-MT2-30B-A3B, a 30B-Parameter Translation Model with 3B Active Parameters Tencent, translation model, mixture-of-experts, Hy-MT2, Tencent Cloud, multilingual AI https://tpsreport.news/news/adobe-firefly-audio-tools-generally-available TPS — Tokens Per Second en 2026-08-20T13:05:45.821Z Adobe Firefly's Generate Music, Speech, and Sound Effects Tools Now Generally Available Adobe, Firefly, AI music generation, voice generation, sound effects, ElevenLabs, Gemini Omni Flash https://tpsreport.news/news/gen-1-5-generalist-ai-robot-single-demo TPS — Tokens Per Second en 2026-08-20T12:50:51.476Z Generalist AI's GEN-1.5 Learns New Robot Tasks From a Single Demonstration robotics, in-context-learning, Generalist AI, GEN-1.5, imitation-learning https://tpsreport.news/news/flux-upscale-2k-4k-video TPS — Tokens Per Second en 2026-08-20T12:35:45.757Z Black Forest Labs Launches FLUX Upscale, a Dedicated Tool for 4K Video Upscaling FLUX, Black Forest Labs, video upscaling, 4K, product update https://tpsreport.news/news/meta-ai-mac-app-system-wide-dictation TPS — Tokens Per Second en 2026-08-20T12:20:44.394Z Meta AI Launches Mac App With System-Wide Dictation and Screen Context Awareness Meta AI, Mac app, dictation, Muse Spark, business AI, productivity, Meta https://tpsreport.news/news/openai-private-safety-processing-zero-data-retention TPS — Tokens Per Second en 2026-08-20T08:20:42.578Z OpenAI Launches 'Private Safety Processing' to Detect Misuse Without Storing Enterprise Data OpenAI, AI safety, data privacy, enterprise AI, zero data retention, Anthropic https://tpsreport.news/news/jie-tang-glm-5-3-post-training-scaling-law TPS — Tokens Per Second en 2026-08-20T05:35:47.896Z Z.ai CEO Jie Tang: Parameter Count Alone No Longer Predicts Model Capability Z.ai, GLM 5.3, scaling laws, reinforcement learning, Jie Tang, post-training, open-weight models https://tpsreport.news/news/openai-private-safety-processing-vs-anthropic TPS — Tokens Per Second en 2026-08-19T22:21:43.639Z OpenAI Previews 'Private Safety Processing' to Detect Abuse Without Retaining Customer Data OpenAI, Anthropic, privacy, enterprise AI, data retention, AI safety https://tpsreport.news/news/agentcore-web-search-domain-date-filters TPS — Tokens Per Second en 2026-08-19T22:20:44.377Z AWS Adds Runtime Domain and Publish-Date Filters to AgentCore Web Search AWS, Amazon Bedrock AgentCore, Web Search, AI agents, product update, MCP