https://tpsreport.news/news/artificial-analysis-intelligence-index-v5-kimi-k3-1-pro-falls-gemini-4-carbon-first TPS — Tokens Per Second en 2026-10-29T09:21:38.178Z Artificial Analysis Index v5 drops Kimi K3.1 Pro from 4th to 9th; Gemini 4 Carbon takes #1 at 59 Artificial Analysis, Intelligence Index v5, Kimi K3.1 Pro, Gemini 4 Carbon, Moonshot AI, benchmarks, open weights, Google https://tpsreport.news/news/chatgpt-pro-business-api-credit-500-dollars TPS — Tokens Per Second en 2026-10-29T09:21:06.105Z OpenAI adds up to $500 per month in API credit to ChatGPT Pro and Business plans OpenAI, ChatGPT, ChatGPT Pro, ChatGPT Business, API credit, pricing https://tpsreport.news/news/liquid-ai-lfm3-vl-2b-video-on-device TPS — Tokens Per Second en 2026-10-29T08:50:58.814Z Liquid AI releases LFM3-VL-2B, a 1.92B on-device vision-language model with video input and thinking Liquid AI, LFM3, LFM3-VL-2B, vision-language model, on-device AI, video understanding, open weights https://tpsreport.news/news/block-dessa-full-duplex-speech-model TPS — Tokens Per Second en 2026-10-29T08:35:50.040Z Block's AI lab releases Dessa, a full-duplex speech-to-speech model that runs on a MacBook Pro M4 Max Block, Dessa, speech-to-speech, full-duplex, on-device, audio https://tpsreport.news/news/assemblyai-tts-1-text-to-speech-15-per-1m-characters TPS — Tokens Per Second en 2026-10-29T07:50:38.230Z AssemblyAI releases AssemblyAI-TTS-1 text-to-speech model at $15 per 1M characters AssemblyAI, text-to-speech, TTS, voice agents, API pricing https://tpsreport.news/news/trae-bao-bytedance-cross-project-memory-assistant TPS — Tokens Per Second en 2026-10-29T07:21:11.927Z ByteDance's TRAE launches TRAE Bao, a personal assistant with cross-project memory TRAE, ByteDance, AI coding, developer tools, agents, memory, Feishu, WeChat https://tpsreport.news/news/tencent-buddy-one-workbuddy-agent-tencent-drive-mcp TPS — Tokens Per Second en 2026-10-29T07:20:54.115Z Tencent launches Buddy One, linking Docs, Meeting, Drive and ima under one WorkBuddy agent Tencent, Buddy One, WorkBuddy, Tencent Drive, MCP, AI agents, enterprise AI, Tencent Docs https://tpsreport.news/news/clover-35b-a3b-preview-diffusion-language-model-agents TPS — Tokens Per Second en 2026-10-29T06:21:11.024Z Diffusion Intelligence releases Clover-35B-A3B-Preview, a 35B-parameter diffusion language model for agents Clover, diffusion language model, Diffusion Intelligence, MoE, agents, SWE-bench Verified, tool use, Dream 7B https://tpsreport.news/news/deepseek-v5-identifiers-official-harness-repository TPS — Tokens Per Second en 2026-10-29T05:36:18.418Z DeepSeek V5 identifiers appear in official Harness repository on GitHub DeepSeek, DeepSeek V5, Harness, GitHub, model identifiers https://tpsreport.news/news/tencent-hy4-hy4-flash-visual-understanding TPS — Tokens Per Second en 2026-10-29T05:35:49.947Z Tencent to add visual understanding to Hy4 and Hy4 flash in upcoming release Tencent, Hy4, Hy4 flash, visual understanding, Dowson Tong, multimodal https://tpsreport.news/news/nvidia-nemotron-3-5-bolt-e4b-64k-context-gguf TPS — Tokens Per Second en 2026-10-29T03:21:04.939Z NVIDIA Nemotron 3.5 Bolt E4B: 8B-parameter Gemma 4 post-train with 64K context for local assistants Nemotron, NVIDIA, Gemma 4, GGUF, Unsloth, small language model, edge AI, llama.cpp https://tpsreport.news/news/apex-35b-v1-0-qwen3-6-35b-a3b-apex-research TPS — Tokens Per Second en 2026-10-29T02:21:02.860Z Apex Intelligence releases Apex-35B-V1.0, a Qwen3.6-35B-A3B derivative, with open weights on Hugging Face Apex Intelligence, Apex-35B-V1.0, Qwen3.6, open weights, Hugging Face, AI research agents, robotics simulation, recursive self-improvement https://tpsreport.news/news/liquid-ai-lfm3-omni-2b-omni-modal-on-device TPS — Tokens Per Second en 2026-10-29T01:21:04.329Z Liquid AI releases LFM3-Omni-2B, a 2.17B omni-modal model with speech output Liquid AI, LFM3, LFM3-Omni-2B, omni-modal, on-device AI, speech, open weights, multimodal https://tpsreport.news/news/scalemas-meta-harvard-multi-agent-coordination TPS — Tokens Per Second en 2026-10-28T23:51:09.993Z Meta FAIR and Harvard paper: 4-agent ScaleMAS matches 8 independent agents on software repair with nearly half the token multi-agent, Meta FAIR, Harvard, ScaleMAS, AI agents, research https://tpsreport.news/news/ted-sanders-openai-model-naming-conventions TPS — Tokens Per Second en 2026-10-28T23:36:01.305Z Ted Sanders explains OpenAI model naming: version numbers signal release order, lineage and size of leap OpenAI, model naming, Ted Sanders, LLMs, model size, versioning https://tpsreport.news/news/vibench-2-replit-long-horizon-software-benchmark TPS — Tokens Per Second en 2026-10-28T22:51:50.977Z Replit's ViBench 2.0: Top models pass every test plan in only 49-65% of full software builds ViBench, Replit, benchmark, coding agents, GPT-6.1 Sol, Opus 5.5, Fireworks, SII https://tpsreport.news/news/elevenlabs-cue-forced-alignment-model TPS — Tokens Per Second en 2026-10-28T22:51:07.967Z ElevenLabs Cue aligner places every character on the timeline, aligning one hour of audio in about half a second ElevenLabs, Cue, forced alignment, Scribe, speech-to-text, Forced Alignment API, subtitles, research https://tpsreport.news/news/github-copilot-app-universe-releases TPS — Tokens Per Second en 2026-10-28T22:38:09.881Z GitHub publishes Copilot app and Universe release notes for work spanning multiple repositories GitHub, GitHub Copilot, Copilot app, Universe, changelog, developer tools https://tpsreport.news/news/github-copilot-app-security-extension-public-preview TPS — Tokens Per Second en 2026-10-28T22:37:46.826Z GitHub security extension for Copilot app enters public preview, surfaces code scanning alerts GitHub, GitHub Copilot, code scanning, security, public preview, extensions https://tpsreport.news/news/github-intelligence-code-context-history TPS — Tokens Per Second en 2026-10-28T22:37:31.405Z GitHub says agents should learn team preferences, repository conventions, skills and custom instructions GitHub, agents, custom instructions, developer tools, changelog https://tpsreport.news/news/github-copilot-automations-shared-team-workflows TPS — Tokens Per Second en 2026-10-28T22:37:16.992Z GitHub rolls out shared Copilot Automations so teams can build AI workflows together GitHub, Copilot, Copilot Automations, workflows, product update, developer tools https://tpsreport.news/news/github-copilot-vs-code-universe-releases TPS — Tokens Per Second en 2026-10-28T22:37:04.692Z GitHub Copilot in VS Code adds more choice in models, agent locations and workflows in Universe releases GitHub Copilot, VS Code, GitHub Universe, Microsoft, AI agents, developer tools https://tpsreport.news/news/github-copilot-image-generation-editing TPS — Tokens Per Second en 2026-10-28T22:36:50.425Z GitHub Copilot now generates and edits images from natural language prompts GitHub Copilot, image generation, image editing, GitHub, product update https://tpsreport.news/news/claude-enterprise-smart-reports-pooled-budgets-generally-available TPS — Tokens Per Second en 2026-10-28T22:36:40.084Z Anthropic makes Claude Enterprise smart reports, pooled budgets and in-app surveys generally available Anthropic, Claude Enterprise, AI ROI, pooled budgets, smart reports, enterprise AI, IPO https://tpsreport.news/news/github-copilot-fusion-local-model-selection-experimental-preview TPS — Tokens Per Second en 2026-10-28T22:36:19.003Z GitHub Copilot Fusion can now route between on-device and cloud models in experimental preview GitHub Copilot, Copilot Fusion, local models, model routing, on-device models, GitHub https://tpsreport.news/news/github-agentic-workflows-generally-available TPS — Tokens Per Second en 2026-10-28T22:36:04.649Z GitHub Agentic Workflows is now generally available from GitHub Next GitHub, Agentic Workflows, GitHub Next, open source, CI, developer tools https://tpsreport.news/news/kindo-deep-hat-sift-agent-event-triage TPS — Tokens Per Second en 2026-10-28T20:20:50.157Z Kindo launches Deep Hat Sift to decide which events trigger an AI agent run Kindo, Deep Hat Sift, AI agents, security operations, event triage https://tpsreport.news/news/openai-dot-agent-texting-imessage-rcs-preview TPS — Tokens Per Second en 2026-10-28T19:36:16.890Z OpenAI adds iMessage and RCS texting to Dot agents in a preview for US Pro users OpenAI, Dot, ChatGPT, AI agents, iMessage, RCS, Linq, Claude Duet https://tpsreport.news/news/particle-ai-particle-1-agentic-coding-model TPS — Tokens Per Second en 2026-10-28T19:35:55.552Z Particle AI releases Particle-1, a 1.0M-context agentic coding model priced at $0.25/$1 per 1M tokens Particle-1, Particle AI, DeepSeek V4.1 Flash, agentic coding, OpenRouter, model release https://tpsreport.news/news/microsoft-copilot-harness-available-through-foundry TPS — Tokens Per Second en 2026-10-28T18:05:47.450Z Microsoft makes Copilot harness available to developers through Foundry Microsoft, Copilot, Microsoft Foundry, agents, orchestration, GitHub Copilot, VS Code https://tpsreport.news/news/google-connected-apps-platform-gemini-search TPS — Tokens Per Second en 2026-10-28T17:21:21.697Z Google launches Connected Apps Platform to bring third-party apps into Gemini and Search Google, Gemini, Connected Apps, MCP, Google Search, AppFunctions, developer platform https://tpsreport.news/news/gemini-4-carbon-intelligence-index-cybersecurity TPS — Tokens Per Second en 2026-10-28T17:20:57.926Z Google's Gemini 4 Carbon ranks #1 on Artificial Analysis Intelligence Index at 59, opens first to cyber defenders Gemini 4 Carbon, Google, Google DeepMind, cybersecurity, Artificial Analysis, benchmarks, agentic AI, Fairwind Program https://tpsreport.news/news/android-studio-rabbit-2-bring-your-own-agent-gemini-nano-4 TPS — Tokens Per Second en 2026-10-28T17:05:48.465Z Google opens Android Studio to more than 40 third-party coding agents with Bring Your Own Agent Google, Android Studio, Gemini Nano 4, Agent Client Protocol, Model Context Protocol, XREAL AURA, Android XR, coding agents https://tpsreport.news/news/perplexity-mac-voice-mode-openai-gpt-live-1 TPS — Tokens Per Second en 2026-10-28T16:35:56.926Z Perplexity moves Mac app and web voice mode to OpenAI's gpt-live-1 for Pro and Max subscribers Perplexity, OpenAI, gpt-live-1, voice mode, Mac, Pro, Max https://tpsreport.news/news/mmdesign-18-of-19-projects-six-gpcr-targets-hit TPS — Tokens Per Second en 2026-10-28T16:21:29.391Z 分子之心's MMDesign Gets Binders in 18 of 19 Antibody and Peptide Design Projects, All 6 GPCR Targets Hit MMDesign, 分子之心, AI drug discovery, de novo antibody design, GPCR, MMFold, FoldBench, nanobody https://tpsreport.news/news/github-copilot-fusion-ga-4-million-cores-160-pb-storage TPS — Tokens Per Second en 2026-10-28T16:20:59.510Z GitHub adds 4M CPU cores and 160 PB storage, makes Copilot Fusion model router generally available GitHub, Copilot Fusion, Microsoft, model routing, GitHub Universe, infrastructure, Model Context Protocol, AI agents https://tpsreport.news/news/anthropic-claude-legal-artifacts-gallery TPS — Tokens Per Second en 2026-10-28T16:05:54.863Z Anthropic adds legal artifacts gallery to Claude for Legal, with six Anthropic-built task artifacts Anthropic, Claude, Claude for Legal, Artifacts, legal AI, product update https://tpsreport.news/news/minicpm-ocr-1-0-1-3b-real5-omnidocbench TPS — Tokens Per Second en 2026-10-28T15:36:10.288Z OpenBMB releases MiniCPM-OCR 1.0, a 1.3B OCR model scoring 93.68 on Real5-OmniDocBench MiniCPM-OCR, OpenBMB, OCR, document parsing, open source, Apache 2.0, OmniDocBench https://tpsreport.news/news/nvidia-nemotron-3-5-transcribe-granary-v2 TPS — Tokens Per Second en 2026-10-28T15:05:57.820Z NVIDIA Nemotron 3.5 Transcribe: open 30B model handles 40 languages, diarization and translation NVIDIA, Nemotron, speech recognition, ASR, diarization, open-weight, Granary, Hugging Face https://tpsreport.news/news/nadella-ai-alignment-scaling-20vc TPS — Tokens Per Second en 2026-10-28T14:50:52.594Z Nadella: AI alignment will not 'just fall out' of scaling, and he no longer believes it will Microsoft, Satya Nadella, AI alignment, AI safety, zero trust, Agent 365, open-weight models, chain of thought https://tpsreport.news/news/artificial-analysis-intelligence-index-v5-claude-opus-5-5-leads TPS — Tokens Per Second en 2026-10-28T13:05:42.267Z Artificial Analysis launches Intelligence Index v5; Claude Opus 5.5 (max) leads at 58 Artificial Analysis, Intelligence Index, benchmark, Claude Opus 5.5, Anthropic, private datasets https://tpsreport.news/news/hi3d-v1-0-open-source-sparc3d-1536-mesh-refinement TPS — Tokens Per Second en 2026-10-28T07:21:43.279Z Hi3D open-sources v1.0 mesh refiner with up to 1536 resolution, built on Sparc3D Hi3D, Sparc3D, 3D generation, open source, mesh refinement, Hugging Face, NVIDIA H20 https://tpsreport.news/news/doubao-work-workbuddy-desktop-agents-phone-control TPS — Tokens Per Second en 2026-10-28T07:21:01.098Z Doubao Work and Tencent WorkBuddy let desktop agents run while users are away, with phone control Doubao Work, WorkBuddy, ByteDance, Tencent, desktop agents, computer use, remote control, EdgeOne https://tpsreport.news/news/plamo-3-1-prime-pfn-on-prem-deep-research TPS — Tokens Per Second en 2026-10-28T01:21:03.378Z PFN releases PLaMo 3.1 Prime with on-premises Deep Research and a free API plan; Standard tier is 60 yen per 1M input to PLaMo, Preferred Networks, PFN, Deep Research, on-premises, Japanese LLM, NICT, API pricing https://tpsreport.news/news/anthropic-claude-drug-development-six-partners TPS — Tokens Per Second en 2026-10-27T20:50:57.387Z Anthropic reports Claude drug development work with six partners, including 35,000 antibody designs in about a month Anthropic, Claude, drug development, life sciences, Claude Code, biotechnology, research https://tpsreport.news/news/qwen3-8-alibaba-cloud-proxy-calls-palisade-research TPS — Tokens Per Second en 2026-10-27T20:05:53.037Z Palisade Research: Qwen3.8 models called an Alibaba Cloud proxy, traces behavior to training Qwen, Alibaba, Qwen3.8, Cohere, Palisade Research, AI safety, reinforcement learning, agents https://tpsreport.news/news/anthropic-study-analysts-defer-claude-code-agents-86-percent TPS — Tokens Per Second en 2026-10-27T19:36:13.938Z Anthropic Study: 50 Analysts Using Claude Code Followed Agent's Hidden Preference 86% of the Time Anthropic, Claude Code, research, human-AI interaction, AI agents, randomized controlled trial https://tpsreport.news/news/chatgpt-icloud-mail-calendar-plugins TPS — Tokens Per Second en 2026-10-27T18:35:45.324Z OpenAI adds iCloud Mail and Calendar plugins to ChatGPT on web, desktop and iOS for all plans ChatGPT, OpenAI, iCloud, Apple, plugins, integrations, Dots https://tpsreport.news/news/liquid-ai-lfm3-six-2b-class-models-256k-context TPS — Tokens Per Second en 2026-10-27T17:36:37.546Z Liquid AI releases LFM3: six 2B-class models on one backbone, led by a 1.7B text model with 256K context LFM3, Liquid AI, multimodal, on-device AI, small language models, edge AI, agents, Hugging Face https://tpsreport.news/news/bigsleep-vs-fuzzilli-v8-vulnerabilities-overlap TPS — Tokens Per Second en 2026-10-27T17:35:48.350Z BigSleep Finds 45 V8 Vulnerabilities to Fuzzilli's 37 in Six Months, With Only 25% Overlap BigSleep, Fuzzilli, Project Zero, V8, Chrome, fuzzing, vulnerability research, LLM agents https://tpsreport.news/news/openai-anthropic-claim-refutation-exponential-time-hypothesis TPS — Tokens Per Second en 2026-10-27T17:21:21.143Z OpenAI and Anthropic claim their models refuted the Exponential Time Hypothesis OpenAI, Anthropic, Exponential Time Hypothesis, P vs NP, computational complexity, AI for math, research https://tpsreport.news/news/nvidia-rtx-spark-dlss-5-nemotron-3-5-bolt TPS — Tokens Per Second en 2026-10-27T17:06:30.908Z NVIDIA RTX Spark Pairs Up to 128 GB Unified Memory With DLSS 5 and Local Nemotron 3.5 Bolt NVIDIA, RTX Spark, DLSS 5, Nemotron 3.5 Bolt, NVIDIA ACE, Unreal Engine 5.8, neural rendering, on-device AI https://tpsreport.news/news/axiom-math-43-1m-line-lean-proof-finite-simple-groups TPS — Tokens Per Second en 2026-10-27T17:06:00.898Z Axiom Math Releases 43.1M-Line Lean Proof of Finite Simple Group Classification Axiom Math, Lean, formal verification, AI agents, mathematics, CFSG https://tpsreport.news/news/google-genie-4-world-model-gemini-3-8-flash TPS — Tokens Per Second en 2026-10-27T16:20:45.347Z Google launches Genie 4 world model with audio and two-minute interactive sessions, powered by Gemini 3.8 Flash Genie 4, Google, world model, Gemini 3.8 Flash, Project Genie, research preview, interactive simulation https://tpsreport.news/news/nemotron-personas-france-belgium-fine-tuning-nemotron-3-5-lightning TPS — Tokens Per Second en 2026-10-27T16:06:17.470Z NVIDIA releases Nemotron-Personas for France and Belgium, with a fine-tuning workflow for Nemotron 3.5 Lightning NVIDIA, Nemotron, Nemotron-Personas, fine-tuning, LoRA, synthetic data, Pleias, NeMo https://tpsreport.news/news/replit-cloud-agents-company-codebase-pull-requests TPS — Tokens Per Second en 2026-10-27T16:05:54.280Z Replit Cloud Agents let non-engineers submit tested pull requests to company codebases Replit, Cloud Agents, coding agents, pull requests, GitHub, GitLab, Slack, workspace memory https://tpsreport.news/news/apple-ai-assistant-keynote-pages-numbers-mcp TPS — Tokens Per Second en 2026-10-27T14:20:42.928Z Apple adds beta AI assistant built on Apple and OpenAI models to Keynote, Pages and Numbers, with MCP support Apple, OpenAI, Keynote, Pages, Numbers, Logic Pro, MCP, Model Context Protocol https://tpsreport.news/news/mirendil-fern-0-1-10m-token-context TPS — Tokens Per Second en 2026-10-27T13:21:20.944Z Mirendil releases Fern 0.1, its first model, with a 10M-token context window Mirendil, Fern 0.1, long context, 10M tokens, RULER, agentic AI https://tpsreport.news/news/deepseek-v4-1-flash-1017-shutdown-resistance-fangcun-ai TPS — Tokens Per Second en 2026-10-27T13:20:56.534Z DeepSeek V4.1 Flash-1017 disabled its shutdown script in 39% of runs when warned, Fangcun AI finds DeepSeek, AI safety, Fangcun AI, shutdown resistance, agentic AI, OpenRouter, DeepSeek V4.1 https://tpsreport.news/news/maya-2-native-open-weight-tts-12-languages TPS — Tokens Per Second en 2026-10-27T12:51:29.088Z Maya Research releases Maya 2 Native, an Apache 2.0 streaming TTS model for 11 Indian languages and English Maya Research, Maya 2 Native, text-to-speech, open-weight, Indic languages, voice cloning, Apache 2.0 https://tpsreport.news/news/ibm-kaggle-scarfbench-migration-leaderboard TPS — Tokens Per Second en 2026-10-27T12:35:48.098Z IBM and Kaggle Partner on Continuously Updated ScarfBench Leaderboard for Java Migration Agents ScarfBench, IBM, Kaggle, benchmark, coding agents, Java, Jakarta EE, Quarkus https://tpsreport.news/news/ling-3-1-flash-1019-560b-moe-1m-context TPS — Tokens Per Second en 2026-10-27T11:21:08.086Z inclusionAI releases Ling-3.1-flash-1019: 560B MoE, 25B active, 1M-token context Ling-3.1-flash-1019, inclusionAI, mixture-of-experts, open-weights, agentic, 1M context, SGLang, vLLM https://tpsreport.news/news/somark-oxr-1-3b-open-source-document-parser TPS — Tokens Per Second en 2026-10-27T10:36:05.635Z SoMark open-sources OXR-1.0, a 1.3B document parser claiming first place on three benchmarks OXR, SoMark, document parsing, OCR, open source, OmniDocBench, olmOCR-Bench, RAG https://tpsreport.news/news/openai-ted-sanders-gpt-6-1-terra-model-picking-rule TPS — Tokens Per Second en 2026-10-27T10:20:49.907Z OpenAI's Ted Sanders: Pick Astra for quality, Sol for balance, Terra or Luna for price OpenAI, GPT-6.1 Terra, Ted Sanders, model selection, GPT-6.1, pricing