https://tpsreport.news/news/flux-3-video-part-1-generation TPS — Tokens Per Second en 2026-08-04T17:36:13.703Z Black Forest Labs Announces FLUX 3 Video, First Details Cover Generation Capabilities FLUX 3, Black Forest Labs, video generation, text-to-video, AI video model https://tpsreport.news/news/chatgpt-atlas-browser-shutdown-august-9 TPS — Tokens Per Second en 2026-08-04T16:21:01.684Z OpenAI to Shut Down ChatGPT Atlas Browser on August 9, Shifts Focus to ChatGPT Desktop App ChatGPT Atlas, OpenAI, browser, product discontinuation, ChatGPT desktop app, AI browser https://tpsreport.news/news/lfm2-5-2-6b-liquid-ai-local-agents TPS — Tokens Per Second en 2026-08-04T14:07:15.871Z Liquid AI Releases LFM2.5-2.6B, a 2.6B-Parameter Agent Model for On-Device Deployment Liquid AI, LFM2.5, on-device AI, small language models, agentic AI, tool calling, open weights, edge AI https://tpsreport.news/news/mistral-shieldstral-3b-safety-classifier TPS — Tokens Per Second en 2026-08-04T14:06:36.987Z Mistral Releases Shieldstral, a 3B Open-Weights Safety Classifier That Matches Models 7x Its Size Mistral, content moderation, open weights, safety classifier, multimodal, Apache 2.0, guardrails https://tpsreport.news/news/lg-k-exaone-2-0-750b-a37b-release TPS — Tokens Per Second en 2026-08-04T05:21:13.267Z LG AI Research Releases K-EXAONE 2.0, a 750B-Parameter Open-Weight MoE Model with 262K Context LG AI Research, K-EXAONE, open-weight models, mixture-of-experts, long-context, multilingual AI, Apache 2.0 https://tpsreport.news/news/qwen-38-max-24t-open-weights-coding TPS — Tokens Per Second en 2026-08-04T04:06:10.677Z Alibaba Unveils Qwen3.8-Max, a 2.4T-Parameter Open-Weight Model for Coding and Agentic Work Qwen, Alibaba, open-weights, coding models, agentic AI, MoE, large language models https://tpsreport.news/news/openai-python-sdk-2-53-0-gpt-5-5-reference TPS — Tokens Per Second en 2026-08-03T21:50:53.550Z OpenAI Python SDK v2.53.0 Adds Support for Unannounced 'GPT-5.5' Model OpenAI, GPT-5.5, Python SDK, Responses API, changelog, SDK update https://tpsreport.news/news/gemini-spark-chrome-web-browsing TPS — Tokens Per Second en 2026-08-03T17:35:56.461Z Google's Gemini Spark Gains Chrome Web-Browsing to Automate Flight Bookings and Errands Google, Gemini Spark, Chrome, AI agents, browser automation, prompt injection https://tpsreport.news/news/qwen-3-8-alibaba-marketing-positive-spin TPS — Tokens Per Second en 2026-08-03T17:20:57.108Z Alibaba Markets Qwen 3.8 as a Job Enhancer, Not a Job Killer — But Skips the Technical Specs Qwen, Alibaba, AI marketing, job automation, labor market https://tpsreport.news/news/amazon-bedrock-automated-reasoning-policy-refinement TPS — Tokens Per Second en 2026-08-03T16:50:50.482Z AWS Adds Automated Policy Refinement to Amazon Bedrock Guardrails' Reasoning Checks Amazon Bedrock, AWS, Guardrails, Automated Reasoning, AI Safety, Formal Verification https://tpsreport.news/news/qwen3-8-max-multimodal-reasoning-model TPS — Tokens Per Second en 2026-08-03T16:05:47.411Z Alibaba Releases Qwen3.8 Max, a Multimodal Reasoning Model with 1M Token Context Qwen, Alibaba, Qwen3.8 Max, multimodal AI, long context, OpenRouter, reasoning model https://tpsreport.news/news/minimax-h3-open-video-model-tops-ranking TPS — Tokens Per Second en 2026-08-03T14:05:58.895Z MiniMax H3 Becomes First Open Video Model to Top an AI Video Ranking MiniMax, video generation, open weights, China AI, multimodal AI, ByteDance, Seedance https://tpsreport.news/news/google-gemini-ai-chrome-1072-bugs TPS — Tokens Per Second en 2026-08-03T12:35:53.071Z Google's Gemini-Based AI Agents Found and Fixed 1,072 Chrome Security Bugs in Two Release Cycles Google, Chrome, Gemini, AI security, vulnerability discovery, Project Zero, AI agents https://tpsreport.news/news/qwen3-8-max-2-4-trillion-parameters TPS — Tokens Per Second en 2026-08-03T11:06:32.010Z Alibaba Releases Qwen3.8-Max, a 2.4 Trillion-Parameter Model Built for Multi-Day Autonomous Tasks Qwen3.8-Max, Alibaba, Qwen, open-weight models, agentic AI, long-horizon tasks, Kimi K3, Moonshot AI https://tpsreport.news/news/gpt-5-6-quantum-crypto-problem-two-teams TPS — Tokens Per Second en 2026-08-03T11:05:51.748Z Two Research Teams Independently Solve Same Quantum Crypto Problem Using GPT-5.6, Three Hours Apart GPT-5.6, OpenAI, quantum cryptography, AI in research, arXiv, unclonable encryption https://tpsreport.news/news/minimax-h3-omni-modal-video-audio-model TPS — Tokens Per Second en 2026-08-03T03:06:12.789Z MiniMax Releases H3, a 33B-Parameter Omni-Modal Model That Generates 2K Video With Native Stereo Audio MiniMax, video generation, multimodal AI, open-weight models, audio generation, transformer architecture