Articles tagged "AI業界動向"

Cue by Manus (cue.im), launched with Manus 2.0, gives Agents a phone, email, wallet and PC. Free in invite-only early access. Not Cue AI (cueai.app). Caveats.

Manus 2.0, announced Sept 28, 2026, adds Cascade, Studio (Video Editor, Game Dev) and Cue. No new plans; Cloud Computer is $30/$50 extra. Cautions included.

Manus Studio, the Manus 2.0 desktop app, works with local files and apps and takes commands from your phone. Apple Silicon Mac and Windows, from $20 a month.

Cloudflare is bringing Deno in: Deno Deploy shuts down around April 2027, the Deno runtime gets 1 more year, JSR continues. Who is affected, where to migrate.

OpenAI dots is an always-on GPT-6 Astra agent announced Sep 29, 2026. One dot is included in Pro and Business Premium. Features, safety, pricing, vs Grok Bot.

Fireworks AI's Ember-1 is Kimi K3 post-trained to cut output tokens by ~40%. Covers benchmarks, pricing, API usage, and how it differs from Kimi K3 itself.

MiMo-V2.6 Flash needs ~170-190GB at 4-bit, ~320-350GB at FP8; Pro ~550-600GB at 4-bit; the 9B distill ~6-7GB. Updated Sep 2026: specs, benchmarks, API prices.

WeKnora is an open-source LLM knowledge platform built by Tencent's WeChat team. It turns documents into RAG search, a ReAct agent, and a self-maintaining wiki under an MIT license you can self-host. This guide covers setup, features, and how it compares to alternatives.

Atria Dawn Preview needs an estimated ~400GB at 4-bit, ~780GB at FP8, or ~1,500GB at BF16. A rundown of this 744B MoE model's specs, vendor-reported benchmarks, API options, and comparison with other open MoE models.

Meta's Muse personal AI agent explained: new Mac app vs iPhone/Android, pricing (free tier, Power $20, Max $100), Secure VM approvals, and US-only rollout.

TypeSafe AI's Jev (Sept 15, 2026) is a System One model: typed answers with probabilities, not text. 70-500ms, $0.042/M input tokens, free output. Updated.

Mercury 2.5, Inception's diffusion LLM launched Sept 8, 2026, hits 1,107 tokens/sec at $0.20/M tokens (80% off at launch), 260K context, API-only, no weights.

A technical look at WeatherNext 3, Google DeepMind's weather AI launched Sep 2026, covering direct satellite training for hourly 5km forecasts and API access.

Meta's Muse Spark 1.3 hits DeepSWE 75.4, Terminal-Bench 88.8. Standard ($1.25/$4.25) vs Contributor ($0.10/$0.20) shows a 12-21x gap paid in training data.

Ox Alpha: free, developer-undisclosed stealth model, ~1M-token context, on OpenRouter/OpenCode since Aug 2026. GLM-5.3 talk runs strong. Updated August 2026.

Bloomberg reported Aug 16, 2026 that Stripe agreed to buy OpenRouter for over $7B. Here's what changes for API keys and billing, and how LiteLLM compares.

August 2026: OpenAI's Ultrafast preview runs GPT-5.6 Sol on Cerebras at up to 750 output tokens/sec, 14x standard. Pricing undisclosed, access limited.

OpenAI Daybreak expansion and GPT-5.6-Cyber, explained: Blue vs. Red access tiers, the Sept 2026 hardware key mandate, and the 16-partner access model.

DOE's Genesis Open Models Initiative, announced Aug 7, 2026, builds open-weight AI for science. First model: Genesis-Science-1, co-developed with Arcee AI.

Overview of DeepSeek V4 Preview, released on April 24, 2026: two open-weight Mixture-of-Experts variants (V4-Pro at 1.6T total / 49B active and V4-Flash at 284B / 13B), 1-million-token context, weights on Hugging Face, and rollout via API and chat — based on official information.

Moonshot AI's Kimi K2.6 reached general availability on April 21, 2026. Reported benchmarks: SWE-Bench Pro 58.6 and HLE-Full (with tools) 54.0, leading the open-weight class. Available across Kimi.com, the Kimi app, the API, and Kimi Code CLI; weights on Hugging Face under a Modified MIT License. This is a follow-up update to our earlier comprehensive guide.

Anthropic released Claude Opus 4.7 to general availability on April 16, 2026. This summary covers the official talking points: notable software-engineering gains over Opus 4.6, high-resolution image input up to 2576px / 3.75MP, the new Task Budgets feature for agentic loops, availability across AWS Bedrock / Vertex AI / Microsoft Foundry, and unchanged pricing at $5 / $25 per MTok.

A unified roundup of eight major AI releases that landed between April 16 and 24, 2026 — Claude Opus 4.7, Kimi K2.6 GA, Gemini 3.1 Pro Deep Research / Deep Research Max, Qwen 3.6-27B, GPT-5.5, DeepSeek V4 Preview, the open-sourcing of Japan's government AI Genai, and OpenClaw 2026.4.23 — with a take on what SMBs should think about.

Summary of OpenAI's GPT-5.5 released on April 23, 2026: improvements in coding, computer use, online research, and document/spreadsheet creation; rollout to ChatGPT Plus/Pro/Business/Enterprise and Codex; API availability; and safety updates — all based on official information.