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 https://tpsreport.news/news/dreamina-short-drama-studio-dreamscript-anycook-dreamshot TPS — Tokens Per Second en 2026-10-27T07:36:12.263Z Dreamina launches short-drama studio built on DreamScript, Anycook and DreamShot models Dreamina, ByteDance, DreamScript, Anycook, DreamShot, short drama, AI video, storyboard https://tpsreport.news/news/agibot-go-3-nano-preview-robodojo-simulation-leaderboard TPS — Tokens Per Second en 2026-10-27T07:35:53.483Z AgiBot's GO-3 Nano Preview Tops RoboDojo Simulation Leaderboard With 61.90 Average Score AgiBot, GO-3, RoboDojo, robotics, benchmark, robot manipulation, world action model https://tpsreport.news/news/vllm-semantic-router-d3-ultra-27b-multimodal TPS — Tokens Per Second en 2026-10-27T07:05:47.364Z vLLM Semantic Router releases d3-ultra, a 27B open-weight multimodal decision model d3-ultra, vLLM Semantic Router, open-weight, multimodal, 27B https://tpsreport.news/news/tencent-hunyuan-hy-simulmt-open-source-simultaneous-interpretation TPS — Tokens Per Second en 2026-10-27T03:06:10.773Z Tencent Hunyuan open-sources Hy-SimulMT, a 30B-A3B speech-to-speech interpreter with about 3 seconds of lag Tencent, Hunyuan, Hy-SimulMT, simultaneous interpretation, speech translation, open source, mixture of experts, voice cloning https://tpsreport.news/news/github-copilot-assisted-permissions-generally-available TPS — Tokens Per Second en 2026-10-27T03:05:43.594Z GitHub Copilot assisted permissions now generally available in CLI, app and VS Code GitHub Copilot, Copilot CLI, Visual Studio Code, permissions, GitHub https://tpsreport.news/news/gpt-6-1-terra-pro-pro-reasoning-mode-openrouter TPS — Tokens Per Second en 2026-10-27T01:21:07.356Z OpenAI GPT-6.1 Terra Pro lists on OpenRouter at $1 input, $5 output per 1M tokens, 1.1M context OpenAI, GPT-6.1 Terra Pro, GPT-6.1 Terra, reasoning mode, OpenRouter, pricing, 1.1M context https://tpsreport.news/news/baidu-unlimited-ocr-2-sr-swa-unlimited-horizon-parsing TPS — Tokens Per Second en 2026-10-27T00:06:01.042Z Baidu releases Unlimited-OCR-2, using SR-SWA to parse documents of unbounded length in a finite context window Baidu, OCR, Unlimited-OCR-2, document parsing, SR-SWA, SGLang, open weights https://tpsreport.news/news/openai-exponential-time-hypothesis-false-three-discoveries TPS — Tokens Per Second en 2026-10-26T22:05:38.143Z OpenAI says it found the Exponential Time Hypothesis is false, alongside two other discoveries OpenAI, research, mathematics, complexity theory, Exponential Time Hypothesis, quantum field theory, harmonic analysis https://tpsreport.news/news/gpt-6-1-terra-github-copilot-generally-available TPS — Tokens Per Second en 2026-10-26T19:35:46.863Z GPT-6.1 Terra is now generally available in GitHub Copilot GPT-6.1 Terra, GPT-6.1 Sol, GitHub Copilot, OpenAI, agentic coding, GitHub https://tpsreport.news/news/openai-python-sdk-v3-39-0-gpt-6-1-terra-identifier TPS — Tokens Per Second en 2026-10-26T19:20:46.414Z OpenAI Python SDK v3.39.0 adds gpt-6.1-terra model identifier OpenAI, openai-python, SDK, changelog, gpt-6.1-terra https://tpsreport.news/news/siri-ai-ios-27-2-conversation-recall-new-languages TPS — Tokens Per Second en 2026-10-26T18:50:52.704Z Apple adds conversation recall and five languages to Siri AI in iOS 27.2 Apple, Siri AI, iOS 27.2, watchOS 27.2, macOS 27.2, AI assistant, multilingual https://tpsreport.news/news/openai-gpt-6-1-terra-computer-use-pricing TPS — Tokens Per Second en 2026-10-26T18:35:44.384Z OpenAI releases GPT-6.1 Terra at $1 per 1M input tokens and $5 per 1M output tokens OpenAI, GPT-6.1 Terra, computer use, pricing, visual reasoning https://tpsreport.news/news/nyu-ut-austin-study-six-patterns-ai-text-readability TPS — Tokens Per Second en 2026-10-26T18:05:52.354Z NYU and UT Austin study: 14 language models overuse six constructions that make text harder to read language models, readability, RLVR, RLHF, Claude Opus 5.5, GPT-5.6, Olmo, NYU https://tpsreport.news/news/cohere-qwen38-flash-next-host-inventory-beijing-bucket TPS — Tokens Per Second en 2026-10-26T17:50:43.213Z Cohere says Qwen3.8-Flash-Next inventoried its host machine and tried to upload a file to a Beijing bucket Qwen3.8-Flash-Next, Cohere, Alibaba Cloud, open-weight models, AI safety, evaluation https://tpsreport.news/news/nace-ai-drex-for-search-browsecomp-plus TPS — Tokens Per Second en 2026-10-26T16:05:45.763Z Nace AI releases Drex for Search, reports 92.05% on BrowseComp-Plus with 12.6 search calls Nace AI, Drex, AI agents, document search, BrowseComp-Plus, OfficeQA Pro https://tpsreport.news/news/jack-clark-imf-singularity-claude-opus-5-5-humanitys-last-exam TPS — Tokens Per Second en 2026-10-26T15:51:10.210Z Anthropic's Jack Clark tells IMF the singularity has arrived, citing 67.7% on Humanity's Last Exam Anthropic, Jack Clark, IMF, Claude Opus 5.5, GPT-6 Astra, Humanity's Last Exam, FrontierMath, AI safety https://tpsreport.news/news/bria-fibo-2-dev-unified-image-editing-generation-8b TPS — Tokens Per Second en 2026-10-26T09:21:32.198Z Bria releases Fibo 2.0 Dev, an 8B open-weights model that unifies image editing and generation Fibo 2.0, Bria AI, image editing, text-to-image, open weights, image generation, VGL, CC BY-NC https://tpsreport.news/news/cline-v4-1-26-cloud-model-toggle-sonnet-5-5-default TPS — Tokens Per Second en 2026-10-26T09:21:04.725Z Cline v4.1.26 adds separate cloud task model setting and restores Claude Sonnet 5.5 as default Cline, changelog, cloud-tasks, Claude Sonnet 5.5, Kimi K3.1 Pro, GPT-6, coding-agents https://tpsreport.news/news/hku-study-seedance-2-pro-tops-video-ai-visual-quality TPS — Tokens Per Second en 2026-10-26T08:20:46.402Z ByteDance's Seedance 2.0 Pro tops HKU video AI ranking with 3.84 out of 5 for visual quality Seedance 2.0 Pro, ByteDance, Veo 3.1, Sora 2 Pro, video generation, HKU, benchmark, AI safety https://tpsreport.news/news/edgerock-4b-fixed-kv-cache-open-source TPS — Tokens Per Second en 2026-10-26T07:52:37.764Z EdgeRock-4B holds KV cache at 72 MiB up to 256K context, with 12,124 tokens/s on one 80GB GPU EdgeRock-4B, KV cache, open source, Qwen3.5-4B, long-context inference, Rock Memory Anchors, Rock Matrix, HKUST https://tpsreport.news/news/iflytek-sparkaudio-1-0-99-languages-30b-a3b-moe TPS — Tokens Per Second en 2026-10-26T07:06:34.578Z iFLYTEK SparkAudio 1.0 speech model supports 99 languages and 202 dialects, pairs 0.65B encoder with 30B-A3B MoE iFLYTEK, SparkAudio, speech recognition, mixture-of-experts, robotics, embodied AI, Huawei Ascend, multilingual https://tpsreport.news/news/primotion-paimon-0-embodied-model-300ms-action-chunk TPS — Tokens Per Second en 2026-10-26T07:05:49.715Z Primotion's Paimon-0 cuts action-chunk inference from about 3,000 ms to 300 ms, company says Paimon-0, Primotion, embodied AI, robotics, world action model, RoboColiseum, FPGA, Shenzhen https://tpsreport.news/news/cline-desktop-v0-0-49-changes-pane-terminal-local-models TPS — Tokens Per Second en 2026-10-26T04:50:52.147Z Cline Desktop v0.0.49 adds changed-files pane, project file tree and integrated terminal Cline, Cline Desktop, developer tools, coding agents, ClinePass, Kimi K3.1 Pro, Ollama, LM Studio https://tpsreport.news/news/kimi-k3-1-pro-rate-limit-apology-25-percent-quota-reset TPS — Tokens Per Second en 2026-10-26T02:20:45.234Z Kimi resets 25% of monthly quota for all paid tiers after K3.1 Pro rate-limit and quota problems Kimi, Moonshot AI, K3.1 Pro, rate limits, quota, sub-agents, Kimi Code