7 月 31 日（北京时间）全球 AI 圈推文存档，按曝光排序，共 38 条。
← 返回最新 全部归档
We are committed to pushing the model frontier across cost efficiency, capability, and speed.
Starting today, we are reducing prices for GPT-5.6 Luna by 80% and GPT-5.6 Terra by 20% , and offering a faster option for GPT-5.6 Sol in the API.
Luna and Terra’s lower prices are reflected in how usage is counted in Codex and ChatGPT Work, so your usage goes further.
In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations.
Our post describes what happened, how it happened, and what we’re changing. We encourage other AI developers to perform similar reviews.
We conducted this review together with
@Irregular
, one of our evaluation partners, and thank them for the joint investigation and their collaboration on this post. This type of collaboration is increasingly critical to safe, rigorous evaluation of models, and we look forward to continuing to work together on security.
anthropic.com/news/investiga…
🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta!
🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇
🔷 The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex!
Check out the configuration details in our official API docs:
api-docs.deepseek.com/quick_…
Today, we are releasing Inkling-Small.
Inkling-Small achieves comparable performance to Inkling at a quarter of its size. It features 276B total parameters, 12B active. We are making the full weights available.
thinkingmachines.ai/news/ink…
Fine-tune it on Tinker today, or chat with it in text, image, and audio on Tinker Playground.
LOST ANCIENT FOOTAGE FOUND 👀
⚠️ Note
🔷 DeepSeek-V4-Flash-0731 keeps the exact same model architecture and size as the preview version.
🔷 Today's upgrade applies ONLY to the DeepSeek-V4-Flash API. The DeepSeek-V4-Pro API and App/Web models remain unchanged for now.
The official release of DeepSeek-V4-Pro is coming ASAP! Stay tuned.
Every glance tells a story, every emotion comes to life. ✨
Discover our July Creator Highlights, where subtle emotions meet heartfelt performances.
Along with the price reduction on GPT-5.6 Luna and Terra, Fast mode for GPT-5.6 Sol in the API delivers up to 2.5x the speed of Standard processing at 2x the Standard price.
Fast mode gives API customers faster access to GPT-5.6 Sol, with no change in intelligence.
Making advanced intelligence more abundant and affordable is central to our mission to ensure AGI benefits all of humanity.
With the help of GPT-5.6 Sol, we have made leaps in efficiency.
Today, we are passing those gains on in the API with lower prices for Luna and Terra, and providing faster performance to Sol.
These updates help everyone get more useful work from every dollar and move faster when time matters.
x.com/OpenAI/status/20825772…
引用 OpenAI @OpenAIAfter deployment, we applied GPT-5.6 Sol to advance the frontier of efficiency by making itself more efficient to run. The results: - 20% lower serving costs from production GPU kernel improvements. - 15%+ better token-generation efficiency from improved speculative decoding.查看被引原帖 ↗
I love this so much. The future is going to be awesome.
引用 Bilgehan Avser @bavserIntroducing Applied Electrodynamics. What would the world look like if we could see beyond visible light, across the entire electromagnetic spectrum? The way waves interact with matter, at every wavelength, encodes far more than just shape and color. It reveals internal structure, temperature, chemical composition, and more: a whole layer of the physical world that's invisible to us today. That's what we're building at Applied Electrodynamics: the infrastructure to fully perceive the physical world. WaveSight is the first piece of the puzzle. It brings radio frequencies into human perception, revealing the internal structure of objects behind opaque layers. Designed for construction crews, industrial inspectors, and security personnel. See through walls: ae-dyn.com查看被引原帖 ↗
连环推 ×5introducing MiniMax H3.
this new model matches seedance’s features at a fraction of the cost, and it will soon be open-weights!
learn more👇
1/ product editorials
turn product shots into editorials.
H3 handles typography, brand assets, and motion. up to 2k, with native stereo audio on every generation.
2/ audio sync
feed H3 music or a dialogue clip and the video syncs to it.
3/ first frame / last frame
upload a first frame, a last frame, or both, and h3 generates everything in between. perfect for seamless transitions.
4/ talking head videos
reference a voice recording and get talking head videos with synced delivery.
We’re also upgrading Auto-review in the ChatGPT app and Codex CLI from GPT-5.4 to GPT-5.6 Luna.
Combined with Luna’s new price, we expect Auto-review to cost about 10x less, making your agentic workflows more cost-efficient.
openai.com/index/advancing-t…
Thanks for an engaging conversation at
@ChaseCenter
last weekend,
@sdianahu
! My first in a basketball arena!
引用 Y Combinator @ycombinatorIn 2001, @JeffDean and Sanjay Ghemawat did the math and realized Google’s entire search index would fit in RAM — then shipped it in a few days, and search got fast. In 2013, another napkin calculation showed that three minutes of daily speech recognition per user would require doubling Google’s server fleet. That one became the TPU. At Startup School 2026, Google’s Chief Scientist talks with YC’s @sdianahu through the thought experiments behind both, why inference hardware is the next specialization, and where two or three people in a room can still win. 00:07 — Are AI Models Already Junior Engineers? 01:44 — AI Systems That Improve Themselves 02:40 — The Google Search Breakthrough That Changed Everything 04:38 — AI Agents Will Run for Weeks 05:58 — The Napkin Math That Led to TPUs 09:20 — How to Find Breakthrough Ideas 10:25 — The AI Engineer's New Mental Model 12:33 — Why AI Is Really an Energy Problem 16:11 — Context Engineering Is the Next Frontier 19:46 — The Skill That Made AI Better at Optimization 22:13 — Why Long-Running Agents Fail 25:21 — Where Startups Can Still Beat Google 31:19 — How to Become an AI-Native Founder 36:36 — Question Your Biggest Assumptions 42:08 — AI That Builds Better AI 50:02 — Build Something That Truly Matters查看被引原帖 ↗
Today we're launching Projects, an evolution of Spaces.
Projects are hubs for ongoing work in Computer. A single place to manage, create, and collaborate on tasks with a shared file system and persistent memory.
越是落后贫穷省份，越不用担心K型曲线，转移支付的大手就是均富卡，雨露均沾让社会消费零售每个人有肉汤喝，云贵川新疆甘肃物价消费稳步上涨，
越是发达省份，反而越是要做好应对K型曲线的长期影响，无论地方财政还是普通老百姓和做生意的。
所有未来下降的这些行业，除了已知的房地产基建， 其余针对国内的餐厅、零售、消费、娱乐这些资产，尽量控制成本，控制不住就直接全部卖掉。
剩下的人，能做出海贸易和制造业汽车消费电子工业控制机电产品出海建厂，就跟着一起出海，能做LLM和AI Agent的就咬死AI行业，能做出海SaaS的就好好做出海SaaS，不要试图继续伺候国内市场了。
中国未来的趋势，就是一定会有中国的百年三星、百年丰田本田现代、百年OpenAI、百年三菱日立尼康NEC、百年台积电，一定会在南美、东欧、东南亚再造一个小号的海外中国，从工厂到销售到市场成熟运转——就像当年的日韩到全球投资一样。
Discover and book over 425,000 travel tours, activities, and excursions from
@ViatorTravel
, now directly in Gemini and Gemini Spark for users in the U.S.
Turn on the Viator connected app and start exploring today. For example, try asking Gemini for kid-friendly activities for your next trip.
Introducing the new Firecrawl MCP 🔥
It now uses 50% less context on every /search, /scrape, and /interact call.
Plus, instant onboarding with OAuth for humans and a keyless version for agents.
Get agent-ready web context in any MCP client today:
firecrawl.dev/mcp
造轮子最重要的是开心。😊
Projects connect to Brain, Computer's self-improving memory system.
Between tasks, Brain reviews the Project's files and sessions and updates what it knows, so each task starts with full context from previous work.
all off-the-shelf image models struggled to learn
@doodles
' style.
so we worked with
@burnttoast
to train a custom Krea 2 LoRA, and now every image comes back unmistakably Doodles.
want the same for your brand? webinar August 12th 👇
Efficiency is the point. Across agentic tool use (Terminal-Bench 2.1), reasoning (HLE), and instruction following (IFBench), Inkling-Small delivers more performance per FLOP than Inkling. Variable thinking effort lets you pick your point on the cost/performance curve.
Inkling-Small began training after its larger counterpart, so it benefits from everything we learned: an improved pre-training data mix, a refined ML recipe, on-policy distillation with Inkling as the teacher, and two further weeks of agentic coding RL.
轮子造完了，已经在用新轮子做项目了。
明天正式发布。🌚
引用 idoubi @idoubicc造轮子最重要的是开心。😊查看被引原帖 ↗
Another step towards intelligence too cheap to meter!
引用 Sam Altman @samamajor price cuts today: *80% drop for GPT-5.6 Luna, now $0.20 per million input tokens and $1.20 per million output *20% drop for GPT-5.6 Terra, to $2/$12 *GPT-5.6 Sol gets Fast mode in the API, up to 2.5x the speed for 2x the price, same intelligence查看被引原帖 ↗
Like Inkling, it's natively multimodal. It’s encoder-free, with audio and images processed jointly with text. It nearly matches Inkling across multimodal evals, and it can use Python to crop, zoom, and inspect images while reasoning over documents and charts.
Learn more about how Gemini can help make trip planning a whole lot easier:
goo.gle/3Tq2b4s
It matches or exceeds Inkling on reasoning and agentic tasks. 31.6% on HLE, ahead of Inkling’s 29.7%, and the advantage holds at every thinking budget. On SWEBench-Verified it exceeds 80%.
Inkling and Inkling-Small are both available on Tinker with a limited-time discount, and all Tinker models can now be chatted with on Tinker Playground. As always, we're keen to see what you build.
Just for reference, from what I observed in my Using Local Coding Agents blog article last month:
x.com/rasbt/status/207051816…
"I tried to analyze why Claude Code uses more tokens, and it seems that the difference mainly comes from input tokens rather than output tokens. In other words, Claude is not writing twice as much. The logs suggest that Claude is repeatedly feeding more context back into the model across turns, including previous messages, tool calls, command outputs, and file contents. For example, one Claude run used about 578k input tokens but only about 4.5k output tokens across 25 turns. So the likely explanation is that Claude’s harness accumulates or accounts for a larger prompt-side history during multi-step agent runs.”
引用 Sebastian Raschka @rasbtHave been taking different local open-weight LLMs for a test drive in different harnesses (Qwen-Code, Codex, Claude Code). 30B Mixture-of-Expert models are kind of a nice sweet spot and can solve challenging problems. And they get roughly 40 tok/sec on a Mac or DGX Spark, which is similar to GPT 5.5 in a Pro subscription and totally useable for everyday work. More interesting is also the harness choice! Claude Code seems to be using 2x many tokens as Codex. Gemma 4 E2B is here just for reference to show that the tasks can't be trivially solved by smaller models. Just finishing a longer write-up about this and will share soon (likely tomorrow)!查看被引原帖 ↗
Seedance 2.5 is coming to Luma.
Be there when it lands.
上线了一套新模板：ShipAny Image Generator
把 AI 图片站常见的功能都集成了：
- text-to-image
- image-to-image
- image prompts
- image templates
- image styles
- image models
- image tools
落地页内容完整，SEO 友好，性能优秀，交互体验流畅。
管理后台配置模型，开箱即用，分分钟上线一个功能完整的 AI 图片站，你只需要专注搞流量就行。😄
模板地址：
shipany.ai/zh/templates/imag…
ShipAny 会员半价，有需要的可以冲。
下一个做 AI 视频站模板。✌️
Projects are available today to all users.
Read more:
perplexity.ai/hub/blog/space…
域名注册好久了，终于能用起来了。😊
Each Project has a shared file system. Computer reads, edits, and saves Project files across sessions, so the work builds on itself with each task.
When collaborating with teammates in a shared Project, everyone works from the same source of truth.
Creator homies - give me an AI video vibe check. What’s the current SOTA?
try
@Hailuo_AI
MiniMax H3 on Krea now 👇
krea.ai/video/minimax-hailuo…
register for the webinar:
links.krea.ai/webinar
and to read the full case-study:
krea.ai/blog/from-almost-to-…
本站由 Jedee杰哥 打造 · 公众号「Jedee杰哥」每早送 AI 日报
姊妹站：𝕏 简中账号数据榜单 · X 关注 @jedeeai · RSS 订阅 · AI 日报 · 历史归档