https://tpsreport.news/news/perplexity-model-council-computer-platform TPS — Tokens Per Second en 2026-07-28T21:35:36.466Z Perplexity Brings Multi-Model 'Council' Feature to Computer Platform, Lets Users Pick Up to 8 AI Models Perplexity, Model Council, Computer, multi-model AI, agentic AI, product update https://tpsreport.news/news/microsoft-vibevoice-asr-bitnet-cpu-speech-recognition TPS — Tokens Per Second en 2026-07-28T20:20:57.081Z Microsoft Releases VibeVoice-ASR-BitNet: 1.58GB Speech Recognition Model Runs Real-Time on CPU, No GPU Needed speech-recognition, ASR, edge-AI, quantization, BitNet, Microsoft Research, CPU-inference, open-source https://tpsreport.news/news/claude-mythos-preview-cryptographic-vulnerabilities TPS — Tokens Per Second en 2026-07-28T19:20:56.281Z Anthropic's Claude Mythos Preview Discovers New Attacks on AES Encryption and Post-Quantum Signature Scheme HAWK Anthropic, Claude Mythos, cryptography, AES, post-quantum cryptography, AI safety, cryptanalysis https://tpsreport.news/news/olmoearth-platform-geospatial-inference TPS — Tokens Per Second en 2026-07-28T16:35:57.291Z Allen Institute Launches OlmoEarth Platform for Continent-Scale Satellite Inference Ai2, OlmoEarth, geospatial AI, satellite imagery, Earth observation, foundation models, infrastructure https://tpsreport.news/news/amazon-scales-back-nova-ai-models-frontier-research TPS — Tokens Per Second en 2026-07-28T16:20:46.500Z Amazon Reportedly Scales Back Nova AI Models, Shifts Resources to New Frontier Research Team Amazon, Nova, AWS, AGI Lab, Pieter Abbeel, Frontier Model Research, Anthropic, OpenAI https://tpsreport.news/news/liquid-ai-lfm2-5-encoders-cpu-long-context TPS — Tokens Per Second en 2026-07-28T15:20:48.291Z Liquid AI Releases LFM2.5-Encoder Models Claiming 3.7x Faster CPU Inference Than ModernBERT at Long Context Liquid AI, encoder models, LFM2.5, CPU inference, open weights, ModernBERT, text classification, masked language modeling https://tpsreport.news/news/perplexity-personal-computer-windows-ai-agent TPS — Tokens Per Second en 2026-07-28T12:35:54.762Z Perplexity Brings Personal Computer AI Agent Tool to Windows Perplexity, Windows, AI agents, Microsoft, enterprise AI, product launch https://tpsreport.news/news/chatgpt-refuses-author-style-mimicry TPS — Tokens Per Second en 2026-07-28T09:20:40.293Z ChatGPT Now Refuses to Mimic Famous Authors' Writing Styles, Citing Copyright ChatGPT, OpenAI, copyright, content policy, AI litigation, authors https://tpsreport.news/news/cursor-start-india-649-plan TPS — Tokens Per Second en 2026-07-28T04:35:52.016Z Cursor Launches ₹649/Month 'Start' Plan for Indian Developers Cursor, pricing, India, coding-assistant, Grok-4.5, product-update https://tpsreport.news/news/gemini-app-thinking-picker-notifications-update TPS — Tokens Per Second en 2026-07-28T01:05:31.376Z Google Simplifies Gemini App's Thinking Level Picker, Adds Notification Controls Gemini, Google, Gemini app, product update, UI update, Gemini Spark https://tpsreport.news/news/kimi-k3-moonshot-ai-2-8t-open-weight-model TPS — Tokens Per Second en 2026-07-28T00:36:28.889Z Moonshot AI Releases Kimi K3: 2.8T-Parameter Open-Weight Model with 1M-Token Context, Now Available via Unsloth Quantiza Kimi K3, Moonshot AI, Unsloth, open-weight models, mixture-of-experts, long-context, multimodal AI, quantization https://tpsreport.news/news/lm-studio-bionic-kimi-k3-support TPS — Tokens Per Second en 2026-07-28T00:35:40.993Z LM Studio Bionic Adds Moonshot AI's 2.8-Trillion-Parameter Kimi K3 Model LM Studio, Kimi K3, Moonshot AI, open-source models, agentic AI, LM Studio Bionic https://tpsreport.news/news/qwen3-7-flash-1m-context-vision-language TPS — Tokens Per Second en 2026-07-28T00:20:40.319Z Alibaba Launches Qwen3.7 Flash: 1M-Context Vision-Language Model at $0.03/$0.13 per 1M Tokens Qwen, Alibaba, vision-language model, multimodal AI, OpenRouter, long context https://tpsreport.news/news/microsoft-mai-cyber-1-flash-cybersecurity-model-1 TPS — Tokens Per Second en 2026-07-28T00:05:44.291Z Microsoft Unveils MAI-Cyber-1-Flash, Claims Cybersecurity Model Beats Rivals at Half the Cost Microsoft, cybersecurity, AI model, MAI-Cyber-1-Flash, Project Perception, OpenAI, GPT-5.4 https://tpsreport.news/news/kimi-k3-2-8-trillion-parameters-license TPS — Tokens Per Second en 2026-07-27T23:50:40.824Z Moonshot AI Releases Kimi K3 Weights: 2.8 Trillion Parameters, Tighter Commercial License Kimi K3, Moonshot AI, open weight models, LLM pricing, LLM licensing, AI in China, OpenRouter https://tpsreport.news/news/kimi-k3-open-weights-moonshot-ai TPS — Tokens Per Second en 2026-07-27T19:50:44.558Z Moonshot AI Open-Sources Kimi K3 Weights After Model Matched GPT-5.6 Sol on Benchmarks Moonshot AI, Kimi K3, open weights, open source, MoE, China AI, distillation https://tpsreport.news/news/microsoft-mai-cyber-1-flash-cybersecurity-model TPS — Tokens Per Second en 2026-07-27T19:05:48.185Z Microsoft Launches MAI-Cyber-1-Flash Security Model, Still Routes Hard Cases to OpenAI's GPT-5.4 Microsoft, OpenAI, cybersecurity, MAI-Cyber-1-Flash, MDASH, GPT-5.4, CyberGym, AI security https://tpsreport.news/news/microsoft-mai-cyber-1-flash-perception TPS — Tokens Per Second en 2026-07-27T18:50:46.600Z Microsoft Launches MAI-Cyber-1-Flash, Its First Cybersecurity Model, With Agentic Security Platform Perception Microsoft, cybersecurity, AI agents, MAI-Cyber-1-Flash, Perception, Mustafa Suleyman, model release https://tpsreport.news/news/china-open-weight-ai-models-kimi-k3-pressure TPS — Tokens Per Second en 2026-07-27T17:05:56.301Z Moonshot AI's Free Kimi K3 Model Is Forcing OpenAI, Google, and Anthropic to Rethink Their Open-Weight Strategy open-weight AI, China AI, Moonshot AI, Kimi K3, OpenAI, Anthropic, Google DeepMind, AI policy https://tpsreport.news/news/threads-meta-ai-private-dms TPS — Tokens Per Second en 2026-07-27T16:35:32.843Z Meta AI Comes to Threads Direct Messages, Ending Public-Feed-Only Chatbot Interactions Meta AI, Threads, product update, messaging, chatbot https://tpsreport.news/news/guardoc-health-amazon-nova-medical-documents TPS — Tokens Per Second en 2026-07-27T16:20:37.093Z Guardoc Health Cuts Documentation Errors 46% Using Amazon Nova Models on Bedrock Amazon Nova, Amazon Bedrock, healthcare AI, document processing, AWS, multimodal AI, Guardoc Health https://tpsreport.news/news/kimi-k3-2-8t-parameter-open-model TPS — Tokens Per Second en 2026-07-27T15:20:59.828Z Moonshot AI Releases Kimi K3: Open-Weight 2.8T-Parameter Model With 1M-Token Context and Native Multimodality Kimi K3, Moonshot AI, open-weight models, mixture-of-experts, long-context, multimodal AI, agentic AI https://tpsreport.news/news/claude-opus-4-7-mirrorcode-robotics-benchmark TPS — Tokens Per Second en 2026-07-27T13:51:03.984Z Anthropic's Claude Opus 4.7 Completes Robot Tasks 20x Faster Than Prior Model, New Benchmark Shows Week-Long Coding Feat Anthropic, Claude Opus 4.7, robotics, MirrorCode, METR, Epoch AI, benchmark, coding agents https://tpsreport.news/news/kilo-code-jetbrains-7-0-11-update TPS — Tokens Per Second en 2026-07-27T13:50:36.721Z Kilo Code Ships JetBrains 7.0.11 with Offline CLI Bundle and Claude Opus/Sonnet 5+ Adaptive Thinking Support Kilo Code, JetBrains, IDE plugin, Claude Opus, Claude Sonnet, Anthropic, AWS Bedrock, changelog https://tpsreport.news/news/metr-expenditure-horizon-ai-agent-cost-metric TPS — Tokens Per Second en 2026-07-27T12:35:49.281Z METR Proposes 'Expenditure Horizon' Metric to Price AI Agents Against Human Labor METR, AI agents, benchmarks, AI research, NanoGPT, Anthropic, OpenAI, ARC-AGI https://tpsreport.news/news/claude-shared-chats-search-engine-exposure TPS — Tokens Per Second en 2026-07-27T08:05:37.992Z Shared Claude Chats Exposed to Search Engines Due to Missing noindex Tag Claude, Anthropic, privacy, data exposure, search engines