https://tpsreport.news/news/duo-1-0-causal-base-inclusionai-full-duplex-world-model TPS: Tokens Per Second en 2026-11-13T09:35:52.944Z inclusionAI releases Duo 1.0 causal-base, a full-duplex multimodal world model initialized from MiniMax H3 Duo, inclusionAI, world model, full-duplex, multimodal, open weights, Hugging Face, MiniMax H3 https://tpsreport.news/news/fx-bio-1031-endless-frontier-open-weight-bioinformatics-model TPS: Tokens Per Second en 2026-11-13T08:22:43.942Z Endless Frontier open-sources Fx-Bio-1031, a 125B MoE bioinformatics model with 512K context Fx-Bio-1031, Endless Frontier, bioinformatics, open-weight, MoE, BioMysteryBench, Qwen, RSI https://tpsreport.news/news/naive-n0-5-flash-latent-space-test-2100-tokens-per-second TPS: Tokens Per Second en 2026-11-13T08:21:30.210Z Latent Space test: NaiveAI's Naive-N0.5-Flash hits about 2,100 tokens/s, 3.1s median response Naive-N0.5-Flash, NaiveAI, Latent Space, DeepSeek V4.1 Flash, GLM 5.3 Flash, NL2Repo-Bench, open-source, inference speed https://tpsreport.news/news/huawei-cloud-cloudrobo-wm1-2-ascend-native-world-model-nov-16 TPS: Tokens Per Second en 2026-11-13T02:50:47.820Z Huawei Cloud to unveil CloudRobo-WM1.2, its first Ascend-native world model, on November 16 Huawei Cloud, CloudRobo-WM1.2, world model, embodied AI, Ascend, robotics, Industry AI Dream Factory https://tpsreport.news/news/liquid-ai-liquidcodec-74m-streaming-audio-codec TPS: Tokens Per Second en 2026-11-12T23:05:53.004Z Liquid AI releases LiquidCodec, a 74M-parameter streaming audio codec running at 12.5 Hz with 12 codebooks Liquid AI, LiquidCodec, audio codec, LFM3-Audio-2B, LFM3-Omni-2B, streaming audio https://tpsreport.news/news/blackwell-fingerprints-chinese-open-weight-models-deepseek-kimi-minimax-ling TPS: Tokens Per Second en 2026-11-12T22:51:10.984Z Analysis finds Blackwell fingerprints in six open-weight models from four Chinese labs DeepSeek, Moonshot AI, MiniMax, Ant Group, NVIDIA Blackwell, export controls, Kimi K3, DeepSeek-V4 https://tpsreport.news/news/meta-muse-phone-calls-us-businesses-paid-subscribers TPS: Tokens Per Second en 2026-11-12T22:37:27.851Z Meta's Muse Now Places Phone Calls to U.S. Businesses, Paid Subscribers Only Meta, Muse, AI agents, phone calls, consumer AI, Alexandr Wang, Superintelligence Labs https://tpsreport.news/news/gemini-4-hydrogen-google-search-global-rollout TPS: Tokens Per Second en 2026-11-12T22:37:10.051Z Google rolls out Gemini 4 Hydrogen to Search for all users globally Gemini 4 Hydrogen, Google Search, Google, Gemini, AI Mode, query fan-out https://tpsreport.news/news/replit-agent-agentic-commerce-google-ads TPS: Tokens Per Second en 2026-11-12T22:36:43.573Z Replit Agent can now buy supplies via Link and Visa and run Google Ads, with owner approval on spend Replit, Replit Agent, agentic commerce, Google Ads, Stripe Link, Visa, Browserbase, AI agents https://tpsreport.news/news/gemini-4-hydrogen-artificial-analysis-intelligence-index-54 TPS: Tokens Per Second en 2026-11-12T22:35:51.394Z Gemini 4 Hydrogen scores 54 on Artificial Analysis Intelligence Index, just below Claude Sonnet 5.5 and GPT-6.1 Sol Gemini 4 Hydrogen, Google DeepMind, Artificial Analysis, Intelligence Index, Claude Sonnet 5.5, GPT-6.1 Sol, benchmark https://tpsreport.news/news/nvidia-nemotron-pii-japan-personas-finance-japan-datasets TPS: Tokens Per Second en 2026-11-12T22:21:52.728Z NVIDIA releases two synthetic Japanese datasets, Nemotron-Personas-Finance-Japan and Nemotron-PII-Japan NVIDIA, Nemotron, synthetic data, Japanese, PII, Hugging Face, datasets, finance https://tpsreport.news/news/anthropic-claude-meetings-beta-video-calls TPS: Tokens Per Second en 2026-11-12T20:50:44.620Z Anthropic launches Claude Meetings beta, putting Claude in video calls for Team and Enterprise plans Anthropic, Claude, Claude Meetings, video calls, Team plan, Enterprise, beta https://tpsreport.news/news/sg-legalclarify-scale-ai-imda-singapore-legal-benchmark TPS: Tokens Per Second en 2026-11-12T19:51:31.883Z Scale AI and IMDA launch SG-LegalClarify, testing 19 models on 1,860 Singapore law scenarios Scale AI, IMDA, SG-LegalClarify, benchmark, legal AI, Singapore, evaluation https://tpsreport.news/news/gemini-4-hydrogen-orchestrator-antigravity TPS: Tokens Per Second en 2026-11-12T19:50:57.469Z Google Antigravity adds Gemini 4 Hydrogen and Orchestrator, which switches between Hydrogen and Carbon mid-session Gemini 4, Gemini 4 Hydrogen, Gemini 4 Carbon, Google Antigravity, Orchestrator, agentic coding, Terminal-Bench https://tpsreport.news/news/ibm-granite-4-1-llmon-preview-3b-8b TPS: Tokens Per Second en 2026-11-12T19:21:01.325Z IBM releases Granite 4.1 3B and 8B LLMON previews to separate instructions from data IBM, Granite 4.1, LLMON, prompt injection, Mellea, Hugging Face, AI security, structured prompts https://tpsreport.news/news/perplexity-decider-v1-3-text-span-extraction-apache-2 TPS: Tokens Per Second en 2026-11-12T19:06:36.609Z Perplexity releases Decider v1.3 and Decider Flash v1.3 with text-span extraction and Apache-2.0 weights Perplexity, Decider, Apache-2.0, open weights, text-span extraction, classification, named-entity recognition https://tpsreport.news/news/grok-bot-mac-apple-mail-messages-notes-integration TPS: Tokens Per Second en 2026-11-12T19:06:13.267Z Grok Bot for Mac adds Messages, Mail, Notes, Reminders, Calendar, Location and Contacts access Grok Bot, Mac, Apple Mail, Apple Messages, AI agents, integrations, product update https://tpsreport.news/news/anthropic-claude-for-manufacturing-17-skills TPS: Tokens Per Second en 2026-11-12T18:35:54.032Z Anthropic Launches Claude for Manufacturing, a 17-Skill Plugin for Quoting, Quality and Maintenance Anthropic, Claude, Claude for Manufacturing, plugins, manufacturing, enterprise AI https://tpsreport.news/news/gemini-4-hydrogen-1m-context-pricing-benchmarks TPS: Tokens Per Second en 2026-11-12T18:21:04.090Z Google launches Gemini 4 Hydrogen: 1M-token context, $1.50 input and $7.50 output per 1M tokens Gemini 4 Hydrogen, Google, Gemini, Antigravity, model release, pricing, benchmarks https://tpsreport.news/news/anthropic-self-hosted-claude-managed-agents TPS: Tokens Per Second en 2026-11-12T18:05:47.882Z Anthropic launches self-hosted Claude Managed Agents for customer cloud accounts and on-premises data centers Anthropic, Claude Managed Agents, self-hosting, AI agents, enterprise AI, on-premises https://tpsreport.news/news/openai-gpt-6-astra-chain-of-thought-monitorability-research TPS: Tokens Per Second en 2026-11-12T17:05:46.481Z OpenAI research ties shrinking chain-of-thought transparency in GPT-6 Astra to training choices on capability and cost OpenAI, GPT-6 Astra, chain of thought, monitorability, AI safety, reinforcement learning, research https://tpsreport.news/news/pinecone-documents-api-gemini-embedding-2 TPS: Tokens Per Second en 2026-11-12T16:50:47.349Z Pinecone Documents API adds Gemini Embedding 2 at $0.20 per 1M tokens Pinecone, Gemini Embedding 2, Google, Vertex AI, embeddings, vector database, document search https://tpsreport.news/news/anthropic-project-glasswing-vulncheck-217-fixed TPS: Tokens Per Second en 2026-11-12T11:06:13.898Z VulnCheck: Project Glasswing has 37,966 candidate vulnerabilities, 217 marked fixed in public records Anthropic, Project Glasswing, Claude Mythos Preview, VulnCheck, cybersecurity, open source, curl, OSS Scanner https://tpsreport.news/news/geodesic-intelligence-novaatom-lite-biomolecule-model TPS: Tokens Per Second en 2026-11-12T10:21:25.298Z Geodesic Intelligence releases NovaAtom-Lite, a biomolecule model it claims tops state-of-the-art systems NovaAtom-Lite, Geodesic Intelligence, drug discovery, biomolecules, Quanquan Gu, life sciences AI https://tpsreport.news/news/gemini-4-carbon-api-pricing-doubles-dec-1 TPS: Tokens Per Second en 2026-11-12T10:21:05.789Z Google opens Gemini 4 Carbon to paying developers; API rates double to $4 input, $20 output per 1M tokens on Dec. 1 Gemini 4 Carbon, Google, API pricing, Gemini 4 Argon, enterprise AI, reasoning model, implicit caching https://tpsreport.news/news/baidu-ernie-lightning-wizshot-me-qianfan TPS: Tokens Per Second en 2026-11-12T08:51:37.962Z Baidu launches ERNIE Lightning reasoning model and ERNIE-Wizshot-ME video model on Qianfan Baidu, ERNIE, ERNIE Lightning, ERNIE-Wizshot-ME, Qianfan, video generation, reasoning model, Kunlunxin https://tpsreport.news/news/seedance-2-5-flash-bytedance-video-model-0-99-yuan-per-second TPS: Tokens Per Second en 2026-11-12T08:51:11.291Z ByteDance launches Seedance 2.5 Flash video model at about 0.99 yuan per second for 720p, 34% below Seedance 2.5 Seedance 2.5 Flash, ByteDance, Volcano Engine, video generation, Volcano Ark, Jimeng, BytePlus https://tpsreport.news/news/tether-qvac-botpsy-0-8b-open-weight-tool-calling TPS: Tokens Per Second en 2026-11-12T04:21:07.631Z Tether releases QVAC BotPsy-0.8B, an open-weight 0.8B model for tool calls on personal devices Tether, QVAC, BotPsy, open-weight, function calling, on-device AI, Qwen3.5, agentic AI https://tpsreport.news/news/meta-muse-dedicated-email-address-muse-ai TPS: Tokens Per Second en 2026-11-12T04:20:47.478Z Meta's Muse AI agent gets its own @muse.ai email address, paid users first Meta, Muse, AI agent, email, personal assistant, product update https://tpsreport.news/news/tencent-yuanbao-search-standalone-ai-search TPS: Tokens Per Second en 2026-11-12T02:35:56.348Z Tencent launches Yuanbao Search, a standalone AI search product at so.qq.com Tencent, Yuanbao, AI search, QQ Browser, WeChat, so.qq.com https://tpsreport.news/news/claude-fable-5-2-vs-opus-5-5-sonnet-5-5 TPS: Tokens Per Second en 2026-11-11T23:51:07.502Z Claude Fable 5.2 scores 65 on Artificial Analysis index but costs about 2x Opus 5.5 per task Claude Fable 5.2, Anthropic, Claude Opus 5.5, Claude Sonnet 5.5, Artificial Analysis, ITmedia, model comparison https://tpsreport.news/news/claude-enterprise-security-center-inference-hooks-edr TPS: Tokens Per Second en 2026-11-11T22:51:01.634Z Anthropic makes Claude Enterprise inference hooks generally available, adds Security Center and EDR integration in beta Claude Enterprise, Anthropic, enterprise security, DLP, EDR, SSO, Claude Code, Cowork https://tpsreport.news/news/zed-v1-29-1-pre-claude-fable-mythos-5-2-git-branch-mode TPS: Tokens Per Second en 2026-11-11T22:36:11.187Z Zed v1.29.1-pre adds Claude Fable 5.2 and Claude Mythos 5.2 support, plus Git Panel branch mode Zed, Claude Fable 5.2, Claude Mythos 5.2, Anthropic, Git, code editor, release notes https://tpsreport.news/news/gemini-decisions-api-1000-questions-per-request TPS: Tokens Per Second en 2026-11-11T18:51:04.946Z Google launches Gemini Decisions API, handling up to 1,000 independent questions per request Google, Gemini, Gemini Decisions API, Gemini Decision 1, Google AI Studio, multimodal, API https://tpsreport.news/news/gemini-live-apple-carplay-global-rollout TPS: Tokens Per Second en 2026-11-11T17:35:44.922Z Google confirms Gemini app with Gemini Live is coming to Apple CarPlay, rolling out globally from today Gemini, Gemini Live, Apple CarPlay, Google, iOS, voice assistant, product update https://tpsreport.news/news/decagon-alto-voice-agent-conversation-model TPS: Tokens Per Second en 2026-11-11T17:21:17.251Z Decagon Introduces Alto, a Model That Decides When Its Voice Agent Speaks and What It Says Decagon, Alto, voice agents, duplex, customer support, Decagon Labs https://tpsreport.news/news/radical-numerics-omnii-defense-biothreat-edge-model TPS: Tokens Per Second en 2026-11-11T17:20:59.764Z Radical Numerics unveils Omnii Defense, an offline model that scores DNA sequences for pathogen threat Radical Numerics, Omnii Defense, biosecurity, genome language model, edge AI, biodefense https://tpsreport.news/news/salesforce-sfr-autorandd-embedding-v4-80-hours TPS: Tokens Per Second en 2026-11-11T16:51:21.218Z Salesforce's SFR-AutoR&D agents built SFR-Embedding v4 in about 80 hours Salesforce, SFR-AutoR&D, SFR-Embedding, autonomous research, AI agents, embeddings https://tpsreport.news/news/elevenlabs-summit-nyc-interaction-engine-memory-routines TPS: Tokens Per Second en 2026-11-11T16:06:04.083Z ElevenLabs previews Interaction Engine, Memory, Routines and ElevenAI at its New York Summit ElevenLabs, ElevenAgents, Interaction Engine, Scribe v3 Medical, ElevenAI, voice AI, speech to text, fine-tuning https://tpsreport.news/news/alphagenome-2-deepmind-100m-cell-profiles-68-tissues TPS: Tokens Per Second en 2026-11-11T15:20:55.968Z Google DeepMind releases AlphaGenome 2, trained on 100M+ cell profiles across 68 tissues AlphaGenome 2, Google DeepMind, genomics, AI for science, biology, API, non-commercial research https://tpsreport.news/news/databricks-genie-agent-multi-step-tasks-ontology TPS: Tokens Per Second en 2026-11-11T15:06:39.220Z Databricks updates Genie agent to run multi-step business tasks unsupervised, built on a new ontology Databricks, Genie, AI agents, enterprise AI, Unity Gateway, Ali Ghodsi https://tpsreport.news/news/databricks-genie-ontology-general-availability TPS: Tokens Per Second en 2026-11-11T15:06:02.794Z Databricks makes Genie Ontology generally available, claims 91% pass rate in Turing benchmark Databricks, Genie Ontology, Genie, Unity Catalog, enterprise agents, benchmark, MCP https://tpsreport.news/news/anthropic-mythos-amd-gpu-port-three-days-cuda TPS: Tokens Per Second en 2026-11-11T14:36:12.572Z Anthropic claims Mythos ported Claude to AMD GPUs in three days, versus six to twelve months Anthropic, Mythos, Claude, AMD, Nvidia, CUDA, AI chips, compute https://tpsreport.news/news/jetbrains-air-ides-2026-3-coding-agents TPS: Tokens Per Second en 2026-11-11T14:35:46.330Z JetBrains ships Air in IDE version 2026.3, running Codex, Claude, Copilot and other coding agents inside the IDE JetBrains, Air, Junie, coding agents, IDE, ACP, MCP, IntelliJ IDEA https://tpsreport.news/news/valory-olas-predict-qwen3-8-27b-open-weight-forecasting TPS: Tokens Per Second en 2026-11-11T14:21:19.358Z Valory unveils open-weight Olas-Predict-Qwen3.8-27B, scoring 67.3% balanced accuracy on prediction markets Valory, Olas, open-weight, forecasting, prediction markets, GPT-5.6, Qwen3.8 https://tpsreport.news/news/deepl-aure-voice-model-tone-emotion-translation TPS: Tokens Per Second en 2026-11-11T14:20:51.098Z DeepL launches Aure voice model that carries tone, emotion and speaking pace into translated speech DeepL, Aure, DeepL Voice, speech translation, voice model, real-time translation, emotion