# AlphaSignal Full Index

Site map: https://alphasignal.ai/llms.txt

## Recent articles

- Title: Sakana AI's SAIL Triples Robot Success Rates by Searching Before Moving
  URL: https://alphasignal.ai/news/sakana-ai-s-sail-triples-robot-success-rates-by-searching-before-moving
  Summary: Sakana AI and the University of Tokyo show that a vision language model can plan robot motions more reliably by testing and refining candidate trajectories in simulation before touching the real robot.
  Published: 2026-09-27T23:13:23.000Z
  Topics: robotics, reasoning
  Company: Sakana AI

- Title: OpenAI Patches GPT-6 Sol and Luna's Silent Image Understanding Bug
  URL: https://alphasignal.ai/news/openai-patches-gpt-6-sol-and-luna-s-silent-image-understanding-bug
  Summary: OpenAI patched an image encoding bug that was quietly hurting vision quality on GPT-6 Sol and GPT-6 Luna across the API, Codex, and computer use.
  Published: 2026-09-27T16:50:19.000Z
  Topics: api, image
  Company: OpenAI Developers

- Title: MiniMax Quietly Slips M3.1-Flash-Preview Into Its Coding Tool
  URL: https://alphasignal.ai/news/minimax-quietly-slips-m3-1-flash-preview-into-its-coding-tool
  Summary: MiniMax quietly rolled out M3.1-Flash-Preview inside MiniMax Code, a lighter sibling to M3 aimed at fast, cheap daily coding work with tunable reasoning effort.
  Published: 2026-09-27T05:24:55.000Z
  Topics: llms, development
  Company: MiniMax_Agent

- Title: Higgsfield Lets Claude Build Fully Editable After Effects Projects
  URL: https://alphasignal.ai/news/higgsfield-lets-claude-build-fully-editable-after-effects-projects
  Summary: Higgsfield's MCP connector lets Claude drive a full motion design pipeline, from image generation to a finished, editable After Effects project.
  Published: 2026-09-26T13:31:11.000Z
  Topics: agents, video
  Company: Higgsfield AI 🧩

- Title: Google's Gemini 3.8 Brings Programmable Voices and Sends TPUs Into Orbit
  URL: https://alphasignal.ai/news/google-s-gemini-3-8-brings-programmable-voices-and-sends-tpus-into-orbit
  Summary: Google's weekly roundup ships expressive TTS models, real-time video avatars, upgraded Notebook study tools, and a satellite carrying TPUs to orbit.
  Published: 2026-09-25T18:28:00.000Z
  Topics: audio, llms
  Company: Google AI

- Title: Anthropic's Claude Breaks Physics Record With a Nine-Loop Particle Calculation
  URL: https://alphasignal.ai/news/anthropic-s-claude-breaks-physics-record-with-a-nine-loop-particle-calculation
  Summary: Anthropic's Claude autonomously pushed a frontier theoretical physics calculation from eight to nine loops on Claude Science, costing only a few thousand dollars.
  Published: 2026-09-25T17:46:08.000Z
  Topics: llms, reasoning
  Company: Anthropic

- Title: Runway Brings AI Video and Image Generation Directly Into Claude
  URL: https://alphasignal.ai/news/runway-brings-ai-video-and-image-generation-directly-into-claude
  Summary: Runway's hosted MCP server plugs Gen-4.5, Seedance 2.5, Veo 3.1, Kling and other models directly into Claude, ChatGPT, Cursor and Replit with no API key.
  Published: 2026-09-25T17:41:35.000Z
  Topics: agents, video
  Company: Runway

- Title: Exa's Agent Ultra Beats OpenAI and Anthropic at Web Research for 54% Less
  URL: https://alphasignal.ai/news/exa-s-agent-ultra-beats-openai-and-anthropic-at-web-research-for-54-less
  Summary: Exa Agent Ultra orchestrates swarms of subagents to run exhaustive web research, topping WANDR, WideSearch and DeepSearchQA at lower cost than frontier rivals.
  Published: 2026-09-25T16:46:16.000Z
  Topics: agents, api
  Company: Exa

- Title: Replit Acquires Atta to Turn Agent Chat Into a BI Powerhouse
  URL: https://alphasignal.ai/news/replit-acquires-atta-to-turn-agent-chat-into-a-bi-powerhouse
  Summary: Replit is folding Atta's purpose-built charting engine into its agent, letting anyone generate interactive business charts inside chat without writing SQL.
  Published: 2026-09-25T16:01:46.000Z
  Topics: vibe coding, business
  Company: Replit ⠕

- Title: Zed's Delta Beats Codex and Claude Code at Half the Requests
  URL: https://alphasignal.ai/news/zed-s-delta-beats-codex-and-claude-code-at-half-the-requests
  Summary: Zed's new agent harness Delta ties or beats Codex and Claude Code on Terminal-Bench accuracy at 0.80 to 1.12 times the cost per pass.
  Published: 2026-09-25T15:50:26.000Z
  Topics: benchmarks, development
  Company: Zed

- Title: Runway Adds Seedream 5.0 Layers to Split Flat Images Into Editable Parts
  URL: https://alphasignal.ai/news/runway-adds-seedream-5-0-layers-to-split-flat-images-into-editable-parts
  Summary: Runway added a one-click tool that decomposes any image into transparent, editable layers using ByteDance's Seedream 5.0 Layers model, no manual masking required.
  Published: 2026-09-25T14:01:02.000Z
  Topics: image, development
  Company: Runway

- Title: Sarvam Releases Saaras V4 to Transcribe All 22 Indian Languages at Once
  URL: https://alphasignal.ai/news/sarvam-releases-saaras-v4-to-transcribe-all-22-indian-languages-at-once
  Summary: Sarvam's new speech-to-text model hits state-of-the-art on all 22 Indian languages plus seven English benchmarks, with five output modes in one model.
  Published: 2026-09-25T11:49:36.000Z
  Topics: audio, benchmarks
  Company: Sarvam

- Title: Stanford's Terminal-Bench-Science Exposes GPT-6 Astra Leading at 63% on Real Science Tasks
  URL: https://alphasignal.ai/news/stanford-s-terminal-bench-science-exposes-gpt-6-astra-leading-at-63-on-real
  Summary: A new 70-task agentic benchmark tests frontier models on real scientific research workflows, with top scores stuck in the low 60s.
  Published: 2026-09-24T23:30:55.000Z
  Topics: benchmarks, agents
  Company: Artificial Analysis

- Title: Zed 1.21 Lets Developers Run Language Server Commands Without Cursor Context
  URL: https://alphasignal.ai/news/zed-1-21-lets-developers-run-language-server-commands-without-cursor-context
  Summary: Zed 1.21 adds a language server command picker that surfaces custom LSP commands beyond code actions, unlocking workflows like Markdown Oxide daily journals.
  Published: 2026-09-24T20:37:38.000Z
  Topics: development, open source
  Company: Zed

- Title: LM Arena Merges 12 AI Leaderboards Into One Free Dashboard
  URL: https://alphasignal.ai/news/lm-arena-merges-12-ai-leaderboards-into-one-free-dashboard
  Summary: Arena rolls out a redesigned leaderboard hub that unifies live model rankings, category standings, first impressions, and news in one view.
  Published: 2026-09-24T20:12:20.000Z
  Topics: benchmarks, llms
  Company: Arena.ai

- Title: Google's Gemini and Veo Team Up to Generate Coherent 10-Minute Videos
  URL: https://alphasignal.ai/news/google-s-gemini-and-veo-team-up-to-generate-coherent-10-minute-videos
  Summary: Google Research unveils a four-part multi-agent stack that keeps characters, props, and scenes consistent across minutes-long AI video narratives.
  Published: 2026-09-24T19:44:12.000Z
  Topics: video, agents
  Company: Google Research

- Title: Midjourney's V8 Drops Seamless Tiles, Smarter Edits, and Live Style Previews
  URL: https://alphasignal.ai/news/midjourney-s-v8-drops-seamless-tiles-smarter-edits-and-live-style-previews
  Summary: Midjourney ships seamless tiling for V8.1 and V8.2, a non-destructive inpainting model, and real-time generation previews on its alpha site.
  Published: 2026-09-24T19:20:37.000Z
  Topics: image, development
  Company: Midjourney

- Title: Perplexity's Photon Slashes Search Latency by 12x for AI Agents
  URL: https://alphasignal.ai/news/perplexity-s-photon-slashes-search-latency-by-12x-for-ai-agents
  Summary: Perplexity's new Rust-based Photon engine powers a Fast Search API preset that cuts agent task costs 68% while returning results in 230ms at p95.
  Published: 2026-09-24T18:07:51.000Z
  Topics: api, retrieval
  Company: Perplexity

- Title: Krea Agent Now Turns Plain-Language Prompts Into Custom Creative Apps
  URL: https://alphasignal.ai/news/krea-agent-now-turns-plain-language-prompts-into-custom-creative-apps
  Summary: Krea just rolled out custom apps inside its creative agent, letting users spin up bespoke image and video tools with a single prompt.
  Published: 2026-09-24T17:54:25.000Z
  Topics: vibe coding, video
  Company: Krea

- Title: Microsoft's run-assert-eval Cuts AI Agent Cross-Account Failures From 43.8% to Zero
  URL: https://alphasignal.ai/news/microsoft-s-run-assert-eval-cuts-ai-agent-cross-account-failures-from-43-8-to
  Summary: Microsoft's new run-assert-eval skill turns agent risk discovery, evaluation, policy generation, and re-testing into a single VS Code prompt.
  Published: 2026-09-24T17:37:47.000Z
  Topics: agents, security
  Company: Microsoft Developer

- Title: Anthropic Now Charges Developers for Claude's Blocked Safety Refusals
  URL: https://alphasignal.ai/news/anthropic-now-charges-developers-for-claude-s-blocked-safety-refusals
  Summary: Anthropic will start billing for requests its classifiers block in three high-risk categories, citing coordinated abuse and false positive rates below 0.1 percent.
  Published: 2026-09-24T17:11:05.000Z
  Topics: security, business
  Company: ClaudeDevs

- Title: Perplexity Brings Portable Computer to AMD Ryzen AI Max PCs
  URL: https://alphasignal.ai/news/perplexity-brings-portable-computer-to-amd-ryzen-ai-max-pcs
  Summary: Perplexity's local-first agent stack now runs on AMD Ryzen AI Max Series chips, expanding beyond NVIDIA RTX and DGX Spark for on-device workflows.
  Published: 2026-09-24T16:35:30.000Z
  Topics: agents, gpus
  Company: Perplexity

- Title: Sakana AI Recruits LSTM Pioneer Schmidhuber to Lead Self-Improving AI Lab
  URL: https://alphasignal.ai/news/sakana-ai-recruits-lstm-pioneer-schmidhuber-to-lead-self-improving-ai-lab
  Summary: The LSTM co-creator joins Sakana AI as Chief Scientific Advisor to lead a new Recursive Self-Improvement Lab focused on physical world models.
  Published: 2026-09-24T15:49:21.000Z
  Topics: business, llms
  Company: Sakana AI

- Title: Odyssey's Agora-2 Lets 20 Players Share One AI-Generated World
  URL: https://alphasignal.ai/news/odyssey-s-agora-2-lets-20-players-share-one-ai-generated-world
  Summary: Odyssey's Agora-2 simulates a shared Diablo II style world for up to 20 humans and agents in real time, with no game engine underneath.
  Published: 2026-09-24T15:37:36.000Z
  Topics: video, robotics
  Company: Odyssey

- Title: Parallel Web Systems Lets AI Agents Query Crunchbase and PubMed Directly
  URL: https://alphasignal.ai/news/parallel-web-systems-lets-ai-agents-query-crunchbase-and-pubmed-directly
  Summary: Parallel now lets agents pull from specialized third-party sources like Crunchbase, Polymarket, and PubMed through three flexible access models.
  Published: 2026-09-24T15:25:16.000Z
  Topics: agents, retrieval
  Company: Parallel Web Systems

- Title: Liquid AI's LFM2.5-VL DSpark Makes Vision AI 3.13x Faster on Apple Silicon
  URL: https://alphasignal.ai/news/liquid-ai-s-lfm2-5-vl-dspark-makes-vision-ai-3-13x-faster-on-apple-silicon
  Summary: Liquid AI's new DSpark drafter brings speculative decoding to LFM2.5-VL-3B, delivering up to 3.13x faster decoding on Apple Silicon with lossless output.
  Published: 2026-09-24T14:35:22.000Z
  Topics: infra, image
  Company: Liquid AI

- Title: Databricks Buys Row Zero to Kill the Ungoverned Spreadsheet
  URL: https://alphasignal.ai/news/databricks-buys-row-zero-to-kill-the-ungoverned-spreadsheet
  Summary: Databricks folds Row Zero's high-scale spreadsheet into Genie, giving finance and ops teams a governed grid tied directly to live enterprise data.
  Published: 2026-09-24T13:32:32.000Z
  Topics: business, infra
  Company: Databricks

- Title: Sarvam AI's Vision 2.1 Beats Gemini at Reading India's 22 Languages
  URL: https://alphasignal.ai/news/sarvam-ai-s-vision-2-1-beats-gemini-at-reading-india-s-22-languages
  Summary: Sarvam's new OCR model tops both English and Indic benchmarks, adds handwriting recognition and key-value extraction across 22 Indian languages.
  Published: 2026-09-24T13:23:23.000Z
  Topics: image, benchmarks
  Company: Sarvam

- Title: Tencent's Hy-MT2 Brings Offline AI Translation to iPhone in 33 Languages
  URL: https://alphasignal.ai/news/tencent-s-hy-mt2-brings-offline-ai-translation-to-iphone-in-33-languages
  Summary: Tencent's on-device translation app hits iOS with 33 languages, offline voice and photo input, all powered by a 440MB quantized model.
  Published: 2026-09-24T07:15:39.000Z
  Topics: audio, llms
  Company: Tencent Hy

- Title: Meta's Muse Slashes Real-Time Avatar Generation from 120 Model Calls to Two
  URL: https://alphasignal.ai/news/meta-s-muse-slashes-real-time-avatar-generation-from-120-model-calls-to-two
  Summary: Meta's new Muse Realtime Avatar streams expressive video of any character in live conversation, distilling a 120-step diffusion teacher into a 2-step causal student.
  Published: 2026-09-24T05:43:20.000Z
  Topics: video, audio
  Company: AI at Meta

- Title: Tencent Cuts AI Training Time by 29% With a Simple Batch Size Trick
  URL: https://alphasignal.ai/news/tencent-cuts-ai-training-time-by-29-with-a-simple-batch-size-trick
  Summary: Tencent Hunyuan researchers show that retuning the learning rate lets you scale RL batch sizes for 29% faster training and 2.29x generation throughput.
  Published: 2026-09-24T03:15:39.000Z
  Topics: post training, training infra
  Company: Tencent Hy

- Title: Anthropic's Claude Opus 5.5 Tops Coding Benchmarks but Costs $13 per Task
  URL: https://alphasignal.ai/news/anthropic-s-claude-opus-5-5-tops-coding-benchmarks-but-costs-13-per-task
  Summary: Anthropic's newest flagship tops the Coding Agent Index at 66, but burns 37% more tokens per task than its predecessor to get there.
  Published: 2026-09-24T01:24:22.000Z
  Topics: benchmarks, development
  Company: Artificial Analysis

- Title: Ant Group's Ming-Image Tops Open-Weight UI Design Rankings Among 81 Models
  URL: https://alphasignal.ai/news/ant-group-s-ming-image-tops-open-weight-ui-design-rankings-among-81-models
  Summary: Ant Group's inclusionAI drops a 6B MIT-licensed text-to-image model that beats every other open-weights contender on UI mockups, dashboards and posters.
  Published: 2026-09-24T00:26:13.000Z
  Topics: image, open source
  Company: Artificial Analysis

- Title: Anthropic Ships Claude Code Cloud Sessions so Developers Can Code Without a Laptop
  URL: https://alphasignal.ai/news/anthropic-ships-claude-code-cloud-sessions-so-developers-can-code-without-a
  Summary: Anthropic promotes Claude Code cloud sessions to general availability with one-time credits of $100 for Pro and $250 for Max users.
  Published: 2026-09-23T21:23:41.000Z
  Topics: development, api
  Company: ClaudeDevs

- Title: Google's Antigravity SDK Lets Gemma 4 Run Agent Workflows Without the Cloud
  URL: https://alphasignal.ai/news/google-s-antigravity-sdk-lets-gemma-4-run-agent-workflows-without-the-cloud
  Summary: Google's Antigravity SDK now runs agentic workflows entirely offline using Gemma 4 26B and LiteRT, with hybrid cloud-local orchestration for privacy-sensitive code.
  Published: 2026-09-23T20:56:24.000Z
  Topics: agents, open source
  Company: Google Gemma

- Title: Cursor's Rollouts Sends AI Agents to Watch Your Code in Production
  URL: https://alphasignal.ai/news/cursor-s-rollouts-sends-ai-agents-to-watch-your-code-in-production
  Summary: Cursor launches Rollouts, a bot that monitors deploys and reverts regressions, alongside a faster Security Reviewer that acts like an engineer, not a linter.
  Published: 2026-09-23T20:45:01.000Z
  Topics: development, security
  Company: Cursor

- Title: PrismML Runs Bonsai 1.7B on Smart Glasses at 2x the Speed
  URL: https://alphasignal.ai/news/prismml-runs-bonsai-1-7b-on-smart-glasses-at-2x-the-speed
  Summary: PrismML and Qualcomm shrink a 2B vision-language model to 1-bit precision, fitting 4x more parameters into smart glasses at twice the token speed.
  Published: 2026-09-23T20:34:51.000Z
  Topics: gpus, image
  Company: PrismML

- Title: Anthropic's Claude Marketplace Unifies 2,000 Partners Into One Enterprise Hub
  URL: https://alphasignal.ai/news/anthropic-s-claude-marketplace-unifies-2-000-partners-into-one-enterprise-hub
  Summary: Anthropic unifies plugins, third-party agents, and consulting partners into a single storefront where enterprise spend commitments double as a purchasing rail.
  Published: 2026-09-23T19:21:42.000Z
  Topics: agents, api
  Company: Claude

- Title: OpenAI's MentalHealthBench Tests AI on Everyday Stress Beyond Crisis Responses
  URL: https://alphasignal.ai/news/openai-s-mentalhealthbench-tests-ai-on-everyday-stress-beyond-crisis-responses
  Summary: OpenAI released an open benchmark built with 80+ clinicians across 22 countries to evaluate how AI models handle the full range of mental health conversations, not just emergencies.
  Published: 2026-09-23T19:08:41.000Z
  Topics: benchmarks, security
  Company: OpenAI

- Title: Liquid AI Brings Liquid Context to Snapdragon so Agents Share Your Memory
  URL: https://alphasignal.ai/news/liquid-ai-brings-liquid-context-to-snapdragon-so-agents-share-your-memory
  Summary: Liquid AI's on-device context layer now runs on Snapdragon's Hexagon NPU, letting phones and cars build a private user profile that any agent can tap.
  Published: 2026-09-23T19:00:25.000Z
  Topics: agents, gpus
  Company: Liquid AI

- Title: Prime Intellect Opens Prime Sandboxes to Run 30M AI Agent Environments
  URL: https://alphasignal.ai/news/prime-intellect-opens-prime-sandboxes-to-run-30m-ai-agent-environments
  Summary: Prime Intellect launches microVM sandboxes purpose-built for agentic RL, priced at a third of major competitors and scaling to tens of thousands of concurrent environments.
  Published: 2026-09-23T18:23:50.000Z
  Topics: agents, training infra
  Company: Prime Intellect

- Title: Anthropic's Claude Uncovers a Hidden CRISPR-Like System in Virus DNA
  URL: https://alphasignal.ai/news/anthropic-s-claude-uncovers-a-hidden-crispr-like-system-in-virus-dna
  Summary: Anthropic's new molecular biology lab used Claude to surface a mysterious enzyme system in phage DNA that resembles CRISPR's cut-and-paste machinery.
  Published: 2026-09-23T18:18:33.000Z
  Topics: llms, development
  Company: Anthropic

- Title: Cognition Embeds Devin Into Microsoft Teams for 90% of Fortune 500
  URL: https://alphasignal.ai/news/cognition-embeds-devin-into-microsoft-teams-for-90-of-fortune-500
  Summary: Cognition brings Devin into Microsoft Teams as a native chat participant and wires it into Outlook, Calendar, OneDrive, SharePoint, To Do, and Entra ID.
  Published: 2026-09-23T18:14:35.000Z
  Topics: development, business
  Company: Cognition

- Title: Cohere Brings Model Vault to Canada for Enterprises Needing Data Sovereignty
  URL: https://alphasignal.ai/news/cohere-brings-model-vault-to-canada-for-enterprises-needing-data-sovereignty
  Summary: Cohere's single-tenant inference platform now runs on Canadian infrastructure, giving domestic enterprises a way to keep model traffic inside national borders.
  Published: 2026-09-23T18:12:17.000Z
  Topics: business, infra
  Company: Cohere

- Title: LM Studio's Bionic Lets Developers Draw Architecture and Code It Instantly
  URL: https://alphasignal.ai/news/lm-studio-s-bionic-lets-developers-draw-architecture-and-code-it-instantly
  Summary: LM Studio's Bionic agent adds a shared Excalidraw canvas so you can sketch diagrams, wireframes and system designs the agent can read and edit alongside you.
  Published: 2026-09-23T17:55:12.000Z
  Topics: development, vibe coding
  Company: LM Studio

- Title: Black Forest Labs' FLUX 3 Action Beats pi0.5 at Robot Control by 1.66x
  URL: https://alphasignal.ai/news/black-forest-labs-flux-3-action-beats-pi0-5-at-robot-control-by-1-66x
  Summary: Black Forest Labs releases a 7B open-weight world action model that tops RoboLab, cutting parameters by 56% and running up to 3.95x faster.
  Published: 2026-09-23T17:46:26.000Z
  Topics: robotics, open source
  Company: Black Forest Labs

- Title: Factory Launches Droids Teams Plan for Small Developer Groups at $46 per Seat
  URL: https://alphasignal.ai/news/factory-launches-droids-teams-plan-for-small-developer-groups-at-46-per-seat
  Summary: Factory rolls out a self-serve Teams plan with centralized billing, individual Pro rate limits, and a $500 credit for early adopters.
  Published: 2026-09-23T17:34:46.000Z
  Topics: development, business
  Company: Factory

- Title: You.com Adds One API Flag That Boosts Search Accuracy by 20%
  URL: https://alphasignal.ai/news/you-com-adds-one-api-flag-that-boosts-search-accuracy-by-20
  Summary: You.com's Knowledge parameter injects licensed real-time data into web search results, lifting benchmark accuracy from 64.5% to 84.2% with 140ms extra latency.
  Published: 2026-09-23T17:32:48.000Z
  Topics: api, retrieval
  Company: You.com

- Title: Epoch's FAB Benchmark Watched GPT-6 Astra Jump from 28% to 80%
  URL: https://alphasignal.ai/news/epoch-s-fab-benchmark-watched-gpt-6-astra-jump-from-28-to-80
  Summary: Epoch AI's Furniture Assembly Benchmark tests whether vision models can spot mistakes in half-built IKEA furniture. Top scores jumped from 28% to 80%.
  Published: 2026-09-23T17:21:56.000Z
  Topics: benchmarks, image
  Company: Epoch AI

- Title: OpenAI's ChatGPT Work Lets Voice Control Your Emails, Slack, and Calendar
  URL: https://alphasignal.ai/news/openai-s-chatgpt-work-lets-voice-control-your-emails-slack-and-calendar
  Summary: OpenAI wires ChatGPT Voice into email, calendar, and Slack plugins, backs it with the GPT-6 model family, and pushes it into ChatGPT Work across web and mobile.
  Published: 2026-09-23T17:08:34.000Z
  Topics: llms, audio
  Company: OpenAI

- Title: Fish Audio's Drama 3 Lets Developers Direct AI Voice Like a Human Actor
  URL: https://alphasignal.ai/news/fish-audio-s-drama-3-lets-developers-direct-ai-voice-like-a-human-actor
  Summary: Fish Audio's Drama 3 preview brings natural-language voice direction, mid-sentence tone shifts, multi-character scenes, and single-word regeneration to text-to-speech.
  Published: 2026-09-23T17:01:51.000Z
  Topics: audio, api
  Company: Fish Audio

- Title: Microsoft's MAI-Image-2.6 Climbs to No. 2 Faster Than Any Rival
  URL: https://alphasignal.ai/news/microsoft-s-mai-image-2-6-climbs-to-no-2-faster-than-any-rival
  Summary: Microsoft says its MAI-Image family gained 283 Elo in ten months, climbing the LMArena text-to-image board 1.6x faster than any rival lab.
  Published: 2026-09-23T16:54:48.000Z
  Topics: image, benchmarks
  Company: Microsoft AI

- Title: Google's Gemini Now Controls Adobe, Linear and 12 More Apps
  URL: https://alphasignal.ai/news/google-s-gemini-now-controls-adobe-linear-and-12-more-apps
  Summary: Gemini adds 13 new Connected Apps across productivity, creativity, and lifestyle, letting users control Adobe, Linear, Peloton, and more from a single chat.
  Published: 2026-09-23T16:02:59.000Z
  Topics: llms, api
  Company: Google Gemini

- Title: Google Ships Gemini 3.8 Flash TTS With Prompt-Directed Emotional Speech
  URL: https://alphasignal.ai/news/google-ships-gemini-3-8-flash-tts-with-prompt-directed-emotional-speech
  Summary: Google's new Flash TTS models bring 200+ expressive audio tags and multi-language narration to AI Studio and the Gemini API for developers building voice apps.
  Published: 2026-09-23T15:25:55.000Z
  Topics: audio, api
  Company: Google AI Studio

- Title: Adobe Buys Topaz Labs for $340M and Brings AI Upscaling to Photoshop
  URL: https://alphasignal.ai/news/adobe-buys-topaz-labs-for-340m-and-brings-ai-upscaling-to-photoshop
  Summary: Adobe closes its $340 million purchase of the AI upscaling specialist, keeping the Topaz brand alive while folding its models into Firefly and Photoshop.
  Published: 2026-09-23T15:06:10.000Z
  Topics: business, image
  Company: Topaz Labs

- Title: NVIDIA's Nemotron 3 Beats 12 Rivals With a 14.72% Speaker Error Rate
  URL: https://alphasignal.ai/news/nvidia-s-nemotron-3-beats-12-rivals-with-a-14-72-speaker-error-rate
  Summary: NVIDIA's 100M-parameter Nemotron 3 Diarization tracks up to 8 speakers in overlapping conversations, topping Voice Arena's benchmark with 14.72% DER.
  Published: 2026-09-23T15:02:41.000Z
  Topics: audio, open source
  Company: NVIDIA AI

- Title: Alibaba's Qwen Intelligence Splits Phone Automation Across Three Specialized Agents
  URL: https://alphasignal.ai/news/alibaba-s-qwen-intelligence-splits-phone-automation-across-three-specialized
  Summary: Alibaba's Qwen Intelligence launches three mobile agents plus open benchmarks, pushing on-phone task automation past 90% end-to-end success on real devices.
  Published: 2026-09-23T11:50:54.000Z
  Topics: agents, benchmarks
  Company: Qwen

- Title: Alibaba's Qwen-Audio 3.1 Slashes Voice API Prices by up to 95%
  URL: https://alphasignal.ai/news/alibaba-s-qwen-audio-3-1-slashes-voice-api-prices-by-up-to-95
  Summary: Alibaba's Qwen team ships a five-model audio stack covering recognition, synthesis, and real-time conversation, with price cuts up to 95 percent.
  Published: 2026-09-23T09:11:23.000Z
  Topics: audio, api
  Company: Qwen

- Title: Pika Cuts Seedance 2.5 Video Costs by 77% With Cheap Draft Mode
  URL: https://alphasignal.ai/news/pika-cuts-seedance-2-5-video-costs-by-77-with-cheap-draft-mode
  Summary: Pika adds Seedance 2.5 Draft Mode: iterate on 480P clips at ten cents per second, then re-render the winner in 1080P using the same seed.
  Published: 2026-09-23T00:23:45.000Z
  Topics: video, api
  Company: Pika

- Title: xAI's Grok Bot Breaks Into Corporate Networks With Google Workspace Connectors
  URL: https://alphasignal.ai/news/xai-s-grok-bot-breaks-into-corporate-networks-with-google-workspace-connectors
  Summary: xAI ships a batch of updates to its agentic Grok Bot: custom network routing, native Google Workspace connectors, a faster desktop app, and snappier bot responses.
  Published: 2026-09-22T22:57:13.000Z
  Topics: development, api
  Company: Grok Bot

- Title: Perplexity Turns AI Failures Into Training Data, Cutting Tool Errors 21%
  URL: https://alphasignal.ai/news/perplexity-turns-ai-failures-into-training-data-cutting-tool-errors-21
  Summary: Perplexity's post-training method teaches its Computer agent to correct mistakes using hint-guided self-distillation, cutting live tool-call failures by 21.2%.
  Published: 2026-09-22T20:21:54.000Z
  Topics: post training, agents
  Company: Perplexity

- Title: Cartesia's Sonic 3.6 Dominates Eight of Nine Global Voice AI Leaderboards
  URL: https://alphasignal.ai/news/cartesia-s-sonic-3-6-dominates-eight-of-nine-global-voice-ai-leaderboards
  Summary: Artificial Analysis extended its Controlled Voice Arena to nine new languages, revealing sharp performance gaps between models across Hindi, Mandarin, Arabic and more.
  Published: 2026-09-22T20:08:54.000Z
  Topics: audio, benchmarks
  Company: Artificial Analysis

- Title: StepFun's StepAudio 3 ASR Tops Speech Benchmark With 1.7% Error Rate
  URL: https://alphasignal.ai/news/stepfun-s-stepaudio-3-asr-tops-speech-benchmark-with-1-7-error-rate
  Summary: StepFun's new speech-to-text model ties for the top spot on the Artificial Analysis WER Index, dominating conversational audio at premium pricing.
  Published: 2026-09-22T19:50:37.000Z
  Topics: audio, benchmarks
  Company: Artificial Analysis

- Title: Vals AI's Claude Opus 5.5 Agents Proved a Faster Shortest-Path Algorithm
  URL: https://alphasignal.ai/news/vals-ai-s-claude-opus-5-5-agents-proved-a-faster-shortest-path-algorithm
  Summary: Ten Claude Opus 5.5 agents collaborated on a message board for 15 hours to produce C-HD, a Lean-verified shortest-path algorithm that beats published bounds in a specific density regime.
  Published: 2026-09-22T18:50:04.000Z
  Topics: agents, reasoning
  Company: Vals AI

- Title: OpenAI Drops GPT-6 Sol and Luna at 50% Less Than GPT-5.6
  URL: https://alphasignal.ai/news/openai-drops-gpt-6-sol-and-luna-at-50-less-than-gpt-5-6
  Summary: OpenAI expands its GPT-6 lineup with two cheaper, faster siblings to Astra, cutting API prices in half and pushing prompt caching further.
  Published: 2026-09-22T18:12:13.000Z
  Topics: llms, api
  Company: OpenAI

- Title: Anthropic's Opus 5.5 Tops Cursor's Coding Benchmark at 40% Lower Cost
  URL: https://alphasignal.ai/news/anthropic-s-opus-5-5-tops-cursor-s-coding-benchmark-at-40-lower-cost
  Summary: Anthropic's new flagship lands in Cursor with a 57.8% CursorBench Max score and 40% cheaper per-task cost than Opus 5.
  Published: 2026-09-22T17:22:13.000Z
  Topics: development, llms
  Company: Cursor

- Title: Anthropic's Opus 5.5 Cuts Agentic Coding Costs 40% While Running 30% Faster
  URL: https://alphasignal.ai/news/anthropic-s-opus-5-5-cuts-agentic-coding-costs-40-while-running-30-faster
  Summary: Anthropic's new flagship model matches Fable 5.1 quality on most work while cutting costs 40%, with faster output and stronger alignment scores.
  Published: 2026-09-22T16:31:01.000Z
  Topics: llms, infra
  Company: Claude

- Title: Runway Launches DIFFUSE to Connect AI Video Creators With Studios
  URL: https://alphasignal.ai/news/runway-launches-diffuse-to-connect-ai-video-creators-with-studios
  Summary: Runway's new DIFFUSE platform turns its Talent Network into a full hiring marketplace, matching AI-fluent creatives with brands, agencies and studios that need them.
  Published: 2026-09-22T14:29:40.000Z
  Topics: business, development
  Company: Runway

- Title: Moonshot AI Turns Kimi Browser Extension Into a Chat-Driven Web Agent
  URL: https://alphasignal.ai/news/moonshot-ai-turns-kimi-browser-extension-into-a-chat-driven-web-agent
  Summary: Moonshot rebrands Kimi WebBridge as the Kimi Browser Extension, adding a sidebar chat, recordable skills, and cross-agent support that runs locally.
  Published: 2026-09-22T11:00:01.000Z
  Topics: development, vibe coding
  Company: Kimi.ai

- Title: Tencent's IWC-Bench Actually Runs AI Web Apps to Catch Broken Buttons
  URL: https://alphasignal.ai/news/tencent-s-iwc-bench-actually-runs-ai-web-apps-to-catch-broken-buttons
  Summary: Tencent Hunyuan's new interactive benchmark actually runs AI-generated web apps, uses code coverage to hunt broken paths, and matches human taste 85.3% of the time.
  Published: 2026-09-22T08:53:39.000Z
  Topics: development, benchmarks
  Company: Tencent Hy

- Title: Amazon Brings Moonshot AI's Kimi K3 to Bedrock at 1.76x OpenRouter Prices
  URL: https://alphasignal.ai/news/amazon-brings-moonshot-ai-s-kimi-k3-to-bedrock-at-1-76x-openrouter-prices
  Summary: Moonshot's 2.8T-parameter open-weight model lands on AWS with a 1M-token context, native vision, and explicit prompt caching for agent workloads.
  Published: 2026-09-22T03:51:03.000Z
  Topics: api, infra
  Company: Kimi.ai

- Title: Tencent's Hy Image 3.5 Merges Generation and Editing Into one Model
  URL: https://alphasignal.ai/news/tencent-s-hy-image-3-5-merges-generation-and-editing-into-one-model
  Summary: Tencent's Hunyuan team launched a preview of Hy Image 3.5, claiming a 30% win rate boost over 3.0 at $0.024 per image.
  Published: 2026-09-22T02:40:42.000Z
  Topics: image, api
  Company: Tencent Hy

- Title: Artificial Analysis' Benchmark Reveals Gemini 3.1 Flash TTS Beats Sonic 3.6 on Pronunciation
  URL: https://alphasignal.ai/news/artificial-analysis-benchmark-reveals-gemini-3-1-flash-tts-beats-sonic-3-6-on
  Summary: A new benchmark scores text to speech models on how correctly they pronounce tricky words, and voice preference does not predict accuracy.
  Published: 2026-09-22T02:26:36.000Z
  Topics: audio, benchmarks
  Company: Artificial Analysis

- Title: StepFun's Step 5 Preview Matches Top Rivals at One-Third the Cost
  URL: https://alphasignal.ai/news/stepfun-s-step-5-preview-matches-top-rivals-at-one-third-the-cost
  Summary: StepFun's new 600B MoE flagship matches Kimi K3 on the Intelligence Index at roughly one third the cost, with open weights due in October.
  Published: 2026-09-22T01:49:19.000Z
  Topics: llms, benchmarks
  Company: Artificial Analysis

- Title: xAI's Grok 4.7 Beats Claude at Enterprise Analysis but Doubles Your Bill
  URL: https://alphasignal.ai/news/xai-s-grok-4-7-beats-claude-at-enterprise-analysis-but-doubles-your-bill
  Summary: xAI's Grok 4.7 jumps 111 Elo on Artificial Analysis's agentic knowledge work benchmark, closing the gap with Anthropic at roughly half the cost per task.
  Published: 2026-09-21T22:57:32.000Z
  Topics: llms, benchmarks
  Company: Artificial Analysis

- Title: Cognition Adds xAI's Grok 4.7 to Devin but Warns of Scope Creep
  URL: https://alphasignal.ai/news/cognition-adds-xai-s-grok-4-7-to-devin-but-warns-of-scope-creep
  Summary: Cognition ships xAI's newest model in Devin CLI and Desktop, where it excels at hard backend work but over-scopes on simpler tasks.
  Published: 2026-09-21T21:27:53.000Z
  Topics: development, benchmarks
  Company: Cognition

- Title: Xiaomi's MiMo-V2.6-Pro Tops Open-Weight Rankings With a 1T-Parameter Model
  URL: https://alphasignal.ai/news/xiaomi-s-mimo-v2-6-pro-tops-open-weight-rankings-with-a-1t-parameter-model
  Summary: Xiaomi's new 1.02T parameter MoE tops the Artificial Analysis open weights leaderboard while costing pennies per task through aggressive caching and sparse activation.
  Published: 2026-09-21T20:11:19.000Z
  Topics: llms, open source
  Company: Artificial Analysis

- Title: Meta's Petal Cable Will Move 1 Petabit per Second Across the Atlantic
  URL: https://alphasignal.ai/news/meta-s-petal-cable-will-move-1-petabit-per-second-across-the-atlantic
  Summary: Meta's Petal cable will deliver 1 petabit per second across the Atlantic using multi-core fiber, doubling capacity over the most advanced systems.
  Published: 2026-09-21T19:27:56.000Z
  Topics: infra, training infra
  Company: Engineering at Meta

- Title: HeyGen's Code2Video Bench Reveals AI Still Fails at Motion Graphics
  URL: https://alphasignal.ai/news/heygen-s-code2video-bench-reveals-ai-still-fails-at-motion-graphics
  Summary: HeyGen and Kaggle drop a benchmark that asks LLMs to write HTML, CSS, and JS that renders into motion graphics videos, then judges the results against human designers.
  Published: 2026-09-21T18:54:21.000Z
  Topics: benchmarks, video
  Company: Kaggle

- Title: Dots Studio's Dots3-Note Preview Tops Open-Weight AI With 76.8% on ARC-AGI-2
  URL: https://alphasignal.ai/news/dots-studio-s-dots3-note-preview-tops-open-weight-ai-with-76-8-on-arc-agi-2
  Summary: Dots Studio's 280B mixture-of-experts model hits 76.8% on ARC-AGI-2 at eight cents per task, topping the open-weight verified leaderboard.
  Published: 2026-09-21T18:37:59.000Z
  Topics: benchmarks, llms
  Company: ARC Prize

- Title: Cognition's Devin Cloud Lets Developers SSH Into AI Coding Sessions
  URL: https://alphasignal.ai/news/cognition-s-devin-cloud-lets-developers-ssh-into-ai-coding-sessions
  Summary: Cognition wired Devin Cloud into the CLI with session handoffs and full SSH access to the agent's VM, blurring local and remote work.
  Published: 2026-09-21T18:34:45.000Z
  Topics: development, infra
  Company: Cognition

- Title: Liquid AI's LFM2.5 Tops Mobile AI Charts Using Half the Memory
  URL: https://alphasignal.ai/news/liquid-ai-s-lfm2-5-tops-mobile-ai-charts-using-half-the-memory
  Summary: Liquid AI's LFM2.5 family posts Pareto-frontier scores on Artificial Analysis's new mobile benchmark, matching a top 3B model at 40% less memory.
  Published: 2026-09-21T17:14:52.000Z
  Topics: llms, benchmarks
  Company: Liquid AI

- Title: Kyutai's Voice of Reason Thinks While It Speaks, Hitting 77% Math Accuracy
  URL: https://alphasignal.ai/news/kyutai-s-voice-of-reason-thinks-while-it-speaks-hitting-77-math-accuracy
  Summary: Kyutai's Voice of Reason turns GLM-4-Voice into a speech-native math solver, jumping GSM8K accuracy from 27.3% to 77.1% with no added latency.
  Published: 2026-09-21T17:01:39.000Z
  Topics: audio, reasoning
  Company: kyutai

- Title: SGLang Squeezes 78% More AI Throughput on NVIDIA's Blackwell GPUs
  URL: https://alphasignal.ai/news/sglang-squeezes-78-more-ai-throughput-on-nvidia-s-blackwell-gpus
  Summary: SGLang, Qwen, and NVIDIA ship 4-bit NVFP4 KV cache on Blackwell, packing 1.78x more context and boosting long-context decode up to 78%.
  Published: 2026-09-21T16:44:21.000Z
  Topics: infra, gpus
  Company: LMSYS Org

- Title: Krea Opens 60-Plus AI Image and Video Models to Coding Agents
  URL: https://alphasignal.ai/news/krea-opens-60-plus-ai-image-and-video-models-to-coding-agents
  Summary: Krea's hosted MCP server lets Claude, Cursor, Codex, and other agents generate images, video, and upscales from plain prompts.
  Published: 2026-09-21T16:33:31.000Z
  Topics: agents, video
  Company: Krea

- Title: Perplexity Computer Adds MiniMax H3 and Seedance to Build Videos in One Thread
  URL: https://alphasignal.ai/news/perplexity-computer-adds-minimax-h3-and-seedance-to-build-videos-in-one-thread
  Summary: Perplexity Computer now generates finished video clips inline via MiniMax H3 and ByteDance Seedance 2.5, bundled with copy and creative for Pro and Max users.
  Published: 2026-09-21T16:25:45.000Z
  Topics: video, api
  Company: Perplexity

- Title: xAI Ships Grok 4.7 With Longer Agent Runs at Unchanged Prices
  URL: https://alphasignal.ai/news/xai-ships-grok-4-7-with-longer-agent-runs-at-unchanged-prices
  Summary: xAI ships Grok 4.7 with longer deliberation, better self-checking, and stronger safety guardrails, matching Grok 4.6 on price and latency.
  Published: 2026-09-21T16:17:53.000Z
  Topics: llms, api
  Company: SpaceXAI

- Title: Microsoft's RetroChimera Plans Drug Synthesis Routes and Wins 9 of 10 Expert Reviews
  URL: https://alphasignal.ai/news/microsoft-s-retrochimera-plans-drug-synthesis-routes-and-wins-9-of-10-expert
  Summary: Microsoft Research's RetroChimera combines two neural networks to plan chemical syntheses that expert chemists prefer over published literature reactions.
  Published: 2026-09-21T15:32:15.000Z
  Topics: llms, data
  Company: Microsoft Research

- Title: Boston Dynamics Opens Atlas Training Facility Inside Hyundai's Georgia Factory
  URL: https://alphasignal.ai/news/boston-dynamics-opens-atlas-training-facility-inside-hyundai-s-georgia-factory
  Summary: Boston Dynamics opens a dedicated Atlas training center inside Hyundai's Georgia EV plant, targeting 25,000 humanoids deployed across factories.
  Published: 2026-09-21T14:05:20.000Z
  Topics: robotics, business
  Company: Boston Dynamics

- Title: Moonshot AI Ships Kimi Code Desktop With 1M-Token AI Coding Agent
  URL: https://alphasignal.ai/news/moonshot-ai-ships-kimi-code-desktop-with-1m-token-ai-coding-agent
  Summary: Moonshot AI ships a desktop coding agent for macOS and Windows, wrapping the Kimi Code CLI in a workspace with parallel agents and a built-in browser.
  Published: 2026-09-21T12:20:02.000Z
  Topics: development, agents
  Company: Kimi Developers

- Title: Unitree's Dex5-S Brings a 22-DoF Robotic Hand to Developers for $6,500
  URL: https://alphasignal.ai/news/unitree-s-dex5-s-brings-a-22-dof-robotic-hand-to-developers-for-6-500
  Summary: Unitree's new Dex5-S packs 22 backdrivable joints into a human-sized robotic hand starting at $6,500, targeting the dexterity bottleneck holding back humanoids.
  Published: 2026-09-21T11:27:59.000Z
  Topics: robotics, gpus
  Company: Unitree

- Title: Alibaba's Qwen-Image-2.1 Ships Transparent Image Generation Without Extra Tools
  URL: https://alphasignal.ai/news/alibaba-s-qwen-image-2-1-ships-transparent-image-generation-without-extra-tools
  Summary: Alibaba open sources a 7B image model with native RGBA output, 10 reference image editing, and unified generation plus editing capabilities.
  Published: 2026-09-20T13:06:39.000Z
  Topics: image, open source
  Company: Qwen

- Title: Alibaba's Qwen3.8 LiveTranslate Cuts Speech Translation Lag to 2.3 Seconds
  URL: https://alphasignal.ai/news/alibaba-s-qwen3-8-livetranslate-cuts-speech-translation-lag-to-2-3-seconds
  Summary: Alibaba's next-gen simultaneous interpretation model cuts average lag to 2.3 seconds across 60 languages, adds speaker diarization, and clones each voice separately.
  Published: 2026-09-19T07:08:11.000Z
  Topics: audio, llms
  Company: Qwen

- Title: xAI's Grok Imagine Image 2.0 Jumps to No. 4 at 29% of OpenAI's Price
  URL: https://alphasignal.ai/news/xai-s-grok-imagine-image-2-0-jumps-to-no-4-at-29-of-openai-s-price
  Summary: xAI's new image model jumps 14 spots to fourth place on the Artificial Analysis leaderboard, undercutting OpenAI's flagship on price by more than 3x.
  Published: 2026-09-18T23:36:52.000Z
  Topics: image, benchmarks
  Company: Artificial Analysis

- Title: Anthropic and Accenture Bet $2B on Embedded AI Safety Audits
  URL: https://alphasignal.ai/news/anthropic-and-accenture-bet-2b-on-embedded-ai-safety-audits
  Summary: Anthropic taps Accenture's Faculty unit to embed evaluators inside its labs, with each side pledging $1B over five years to build safety oversight capacity.
  Published: 2026-09-18T20:05:03.000Z
  Topics: business, security
  Company: Anthropic

- Title: Vals AI's CUA-Bench Humbles Every Frontier Agent Below 20 Points
  URL: https://alphasignal.ai/news/vals-ai-s-cua-bench-humbles-every-frontier-agent-below-20-points
  Summary: Vals AI's new CUA-Bench pits frontier models against six commercial video games with only pixels in and keystrokes out, and every model scores below 20%.
  Published: 2026-09-18T18:24:18.000Z
  Topics: benchmarks, robotics
  Company: Vals AI

- Title: xAI Ships Grok Voice Transcribe 2.0 With Half the Errors at Same Price
  URL: https://alphasignal.ai/news/xai-ships-grok-voice-transcribe-2-0-with-half-the-errors-at-same-price
  Summary: SpaceXAI's new speech-to-text model doubles accuracy over v1.0, tops the streaming leaderboard, and keeps pricing at ten cents per hour.
  Published: 2026-09-18T17:47:40.000Z
  Topics: audio, api
  Company: SpaceXAI

- Title: Factory Private Brings Autonomous Coding Droids Inside Enterprise Networks
  URL: https://alphasignal.ai/news/factory-private-brings-autonomous-coding-droids-inside-enterprise-networks
  Summary: Factory rolls out on-premises and air-gapped deployment for its Droid coding agents, targeting enterprises that cannot ship source code to a vendor cloud.
  Published: 2026-09-18T17:00:55.000Z
  Topics: agents, business
  Company: Factory

- Title: Bolt Adds DeepSeek V4.1 Flash at 10x Cheaper Than V4 Pro
  URL: https://alphasignal.ai/news/bolt-adds-deepseek-v4-1-flash-at-10x-cheaper-than-v4-pro
  Summary: Bolt adds DeepSeek V4.1 Flash to its Forge open-model lineup, promising 10x the usage of V4 Pro and stronger visual design output.
  Published: 2026-09-18T16:42:20.000Z
  Topics: vibe coding, open source
  Company: bolt.new

- Title: GPT-6 Astra Cracks a Decade-Old Voting Theory Problem Nobody Could Solve
  URL: https://alphasignal.ai/news/gpt-6-astra-cracks-a-decade-old-voting-theory-problem-nobody-could-solve
  Summary: Epoch AI marks the first Major Advance on its unsolved-math benchmark, with GPT-6 Astra driving the proof in an interactive session with three mathematicians.
  Published: 2026-09-18T16:33:09.000Z
  Topics: benchmarks, reasoning
  Company: Epoch AI