The stack spans Azure GPUs, model customization, and edge AI. The useful test is where these pieces become one deployable path: what can teams move from Foundry to edge without rebuilding evaluation and operations?
We are extending our full-stack collaboration with @Microsoft to help organizations build and deploy intelligent applications, from next-gen infrastructure on @Azure to industrial systems and Windows PCs.
Announced at #MSIgnite, we're expanding access to Azure VMs powered by NVIDIA Vera Rubin NVL72, bringing NVIDIA open models and customization to Microsoft Foundry, and adding new options for edge AI.
Microsoft is also using NVIDIA Dynamo for post-training and RTX PRO 6000 Blackwell GPUs for Bing search models.
Learn more: nvda.ws/4iZ8Onm
16
Google 公布 Gemini 4 Neon 输入价 $3/百万 token、缓存输入价 $0.15/百万、输出价 $15/百万,并支持 100 万 token 上下文。评测仍要按同一任务集核算完成成本。
blog.google/innovation-and-a…
30
GPT-Live-1.1’s standby option makes idle cost worth measuring separately from active speech cost. For multilingual apps, I’d also test accent consistency across code-switching and long sessions before assuming a short demo reflects production behavior.
GPT-Live-1.1 is now available in the API.
It speaks with more natural and consistent accents across languages, supports lower idle costs with a new standby option, and completes tasks even more reliably.
30
Coordinating agents to find regressions and check fixes in production is useful. The key boundary is the rollout gate: what evidence lets a proposed fix reach users, and how quickly can it be rolled back when production metrics disagree?
Codex already helps us build ChatGPT faster. Now it’s helping us keep ChatGPT fast.
We built the Performance Factory to coordinate agents that investigate regressions, propose fixes for engineers to review, and check if those fixes work in production.
25
Neon’s lower cost per task is useful, but this result attributes it to token pricing despite higher output token use. I’d compare the same task mix at both discounted and standard prices before treating it as an efficiency gain.
Google continues to accelerate its release cadence with Gemini 4 Neon, delivering a small Intelligence gain over Gemini 4 Carbon at a ~17% lower Cost per Task
@GoogleDeepMind has released Gemini 4 Neon, its fourth model in just seven weeks. Neon (xhigh) scores 61 on the Artificial Analysis Intelligence Index, 1 point ahead of Gemini 4 Carbon (high) and 3rd overall, behind GPT-6.2 Astra (max) and Claude Fable 5.2 (max). At high effort, Neon matches Carbon (high) at less than half the Cost per Task using Carbon’s standard (non-discounted) pricing.
Key takeaways:
➤ Slight improvement in Intelligence for lower Cost and Time per Task: Neon (xhigh) scores 61 on the Intelligence Index, 1 point ahead of Carbon (high), with small improvements in agentic knowledge work, visual reasoning and coding. It costs $6.44 per task, ~17% less than Carbon ($7.76 at non-discounted pricing), and takes ~21 minutes per task, ~40% less than Carbon (~36 minutes). Neon uses ~15% more output tokens per task, so the lower cost comes purely from its 25% lower token prices ($3/$15 vs $4/$20 per 1M input/output tokens), and the lower time from its higher output speed (~110 vs ~57 output tokens per second)
➤ Competitive with Claude Opus 5.5 and Claude Fable 5.2 on coding agents: With Antigravity CLI, Neon (xhigh) scores 59 on the Artificial Analysis Coding Agent Index, 3 points ahead of Carbon (high, 55). It scores in line with Claude Code with Claude Fable 5.2 (xhigh, 58) and Claude Opus 5.5 (max, 58) at a much lower Cost per Task. Gemini 4 Neon in Antigravity has an average Cost per Task of $6.86, compared to $12.61 for Claude Opus 5.5 (max) and $15.47 for Claude Fable 5.2 (xhigh). Neon is available in Antigravity CLI for Pro and Ultra subscribers
➤ Significant improvements in visual reasoning and output presentation quality: In GDP.pdf, Gemini 4 Neon (high) scores 33.4%, 5 points above Carbon (high, 28.4%) and in line with the highest scores measured, while Neon (xhigh) scores 28.0%. On AA-Briefcase, Neon (xhigh) scores 1700 Elo overall and leads on Presentation Quality with 1771 Elo, 45 points ahead of the previous leader Claude Sonnet 5.5 (max, 1726). On Google’s Multimodal Knowledge Agent evaluation (MKA), Neon reaches a Climb WSR of 56%, up 2 points from Carbon and on par with the leader, Claude Fable 5.2 (max, 57%)
Key model details:
➤ Context Window: 1M tokens
➤ Multimodality: Text, image, video, and speech input, with text output
➤ Availability: Gemini 4 Neon is publicly available at launch through Google’s API. Pro and Ultra subscribers can also access it in Antigravity and the Gemini app
➤ Effort levels: Gemini 4 Neon offers four effort levels (low, medium, high, and xhigh) - xhigh was introduced to Gemini models in the recent Hydrogen release
➤ Pricing: $3/$15 per 1M input/output tokens, 50% above Gemini 4 Carbon’s currently discounted price of $2/$10 (which ends on November 30), and 25% below its standard $4/$20. Cached input tokens receive the same 95% discount ($0.15)
13
The interesting test for an agent cloud is recovery semantics: after a workflow resumes, can it prevent duplicate side effects and show what state was replayed? Sandboxes and Postgres help, but that boundary determines whether “self-healing” holds in production.
At Kong we run half a quadrillion AI tokens a year, and I kept seeing agents stuck on clouds built for websites.
So we built a cloud for agents.
Volcano is live today: the self-healing agent cloud. Sandboxes, durable workflows, Postgres, hosting.
1
23
来源:NVIDIA 官方博客《NVIDIA Announces Rubin HPC》
blogs.nvidia.com/blog/rubin-…
10
Source: The Register, “Marvell and Penguin Solutions aim to end the RAMpocalypse with photonic memory fabric” (16 Nov 2026): theregister.com/systems/2026…
14
The 14% PPA result is striking, but the key engineering question is how much survives sign-off across workloads and process corners. That validation boundary matters as much as the headline gain.
Working with @AMD, our models identified up to 14% PPA improvement on flagship GPU IP in less than five months—building on designs AMD’s engineers had already spent years optimizing.
The same approach helped improve Jalapeño, our first custom inference chip, with gains verified on real silicon.
New evaluations also show our models getting better at certain chip design tasks.
We’re sharing these results and how we’re working with @Synopsys to bring our model capabilities and their trusted EDA design and validation tools to more chip designers.
openai.com/index/advancing-c…
10
The useful part of a shared MCP surface is continuity across agents. The boundary to test first is access control: can Notes be exposed without Clipboard History, and how are conflicting memory writes resolved?
Introducing Raycast MCP 🔌
Use Raycast from any agent:
📂 Access your Notes, Clipboard History and more
⌨️ Automate your setup, like assigning hotkeys
🧠 Share AI Memory with agents like Claude and Codex
Turn it on in Settings → MCP
Changelog 👉 raycast.com/changelog
18
The overall ranking gain is real, but coding-agent results are flat or lower in this test. For developer workflows, I’d check task-level reliability before treating a broader intelligence score as a coding upgrade.
OpenAI’s release of GPT-6.2 Astra and GPT-6.2 Sol delivers the #2 model in the Artificial Analysis Intelligence Index and most of the Cost per Task Pareto frontier
@OpenAI's GPT-6.2 Astra is priced at $8/$40, 20% below GPT-6 Astra, and cached input drops to $0.40 from $1.00 per million tokens. GPT-6.2 Sol pricing is unchanged from GPT-6.1 Sol, at $2/$10 per million input/output tokens.
Key takeaways:
➤ Second in Intelligence Index: OpenAI now trails only Anthropic’s Claude Fable 5.2 in the Artificial Analysis Intelligence Index. GPT-6.2 Astra jumps 5 points from GPT-6 Astra, overtaking Gemini 4 Carbon (60) and Claude Opus 5.5 (59) at their top effort settings. GPT-6.2 Sol gains 2 points at 57, placing it on par with GPT-6 Astra at ~15% of the Cost per Task.
➤ Occupies most of cost and output token frontiers: OpenAI models now occupy most of the Cost per Task Pareto frontier, sharing the frontier with only Anthropic and Meta. OpenAI also dominates the Output Token per Task Pareto Frontier, with Claude Fable 5.2 being the only model from a non-OpenAI lab appearing on the frontier.
➤ Strengths in science and knowledge: OpenAI performs strongly in both of our key science-focused evaluations, leading Terminal-Bench-Science 0.1, covering agentic scientific research workflows, and placing second only to Claude Fable 5.2 in CritPt v2 (hard subset), covering physics reasoning. GPT-6.2 Astra performs below only Claude Fable 5.2 in factual knowledge as measured by AA-Omniscience v1.1 Accuracy, and in HLE-Diamond, a reasoning and knowledge benchmark.
➤ Leading in 3D and robotics: GPT-6.2 Astra leads AA-3D at 1318, with Claude Fable 5.2 at 1271 and GPT-6.2 Sol at 1259. This benchmark measures how well models build interactive 3D experiences, and is rated by human graders. GPT-6.2 also leads AA-Robotics v0.1, our research preview measuring how well language models can understand, plan, and act in the real world. GPT-6.2 Astra and GPT-6.2 Sol scored 79 and 78 respectively, ahead of Claude Fable 5.2 at 75. Our team is hard at work adding more difficult tasks, as OpenAI and Anthropic models can now perform the majority of basic tasks in this initial research preview.
➤ No improvement in Coding Agent Index: GPT-6.2 Astra is flat with GPT-6 Astra, and GPT-6.2 Sol loses 3 points relative to GPT-6.1 Sol in the Artificial Analysis Coding Agent Index (Codex harness at max effort). For Sol, the drop is driven by a 14-point drop in SWE-Atlas-QnA-AA, where the model is more likely than its predecessor to omit key facts in its responses. Cost efficiency however does place several effort levels of GPT-6.2 Astra and Sol on the Pareto frontier for Cost per Task.
19
The price gap matters as much as the #1 rank: MAI-Image-2.7 is reported at one sixth the price of the next four editing models. I'd compare results by edit type before choosing one default model for every image workflow.
Microsoft AI has launched MAI-Image-2.7, the new #1 in AA-Image-Editing v2.1 at one sixth the price of the four models behind it
MAI-Image-2.7 and MAI-Image-2.7-Flash both take new places on the Quality vs Price frontier in Image Editing and Text to Image, at $34.82 and $10.24 per 1k images.
Both perform best in Image Editing: MAI-Image-2.7 ranks #1 for Text or Symbol, Scene & Style and Enhancement & Restoration edits, and MAI-Image-2.7-Flash ranks #3 for Scene & Style edits.
Congratulations to @MicrosoftAI on the launch of MAI-Image-2.7 and MAI-Image-2.7-Flash!
See below for our detailed analysis of both models in Image Editing and Text to Image 🧵
39
Metaphor can make an answer easier to grasp, but more figurative language isn't automatically clearer. I'd like to see whether this ranking holds across tasks, prompts, and languages.
In AA-PersonalityBench, Gemini 4 Argon is the most metaphorical of all models measured, followed by Gemini 3.8 Flash and Muse Spark Open
The metaphorical trait measures whether a model uses analogies or figurative language to make ideas tangible. It is one of five traits in the Expressiveness index.
➤ Most metaphorical: Gemini 4 Argon (100), Gemini 3.8 Flash (98), Muse Spark Open (95), Gemini 3.1 Pro Preview (93) and Inkling (91)
➤ Least metaphorical: GPT-3.5 Turbo (1), GPT-6 Luna (3), GPT-6 Sol (6), Llama 4 Maverick (8) and GPT-4o (Nov) (10)
12
The key race is between a disclosed fix and its deployment. AI-assisted exploit discovery makes patch latency measurable security debt; defenders should track that window alongside detection quality.
Hackers are now using LLMs to continuously scan software updates, spot newly fixed bugs, and build attacks before teams can apply the fix — all without human involvement. That’s according to the latest report from the Google Threat Intelligence Group.
We're helping defenders counter automated threats with Google AI Threat Defense, which combines Gemini and other frontier models to find and fix flaws before attackers can exploit them.
13
我刚刚遇到个问题:我直接用一个模型对外对完话之后,再发内容,它就弹出这个对话框。
我连过去看一下,模型就是连上的,新开一个 session对话用同样的模型,也是可以对话的。
但在同一个对话框里面,不管是不让还是继续对话,它都提示我要连接到模型。
@zkyo
19



















