2026-08-31 至 2026-09-06 短讯
本页 49 条：12 条有单独页面，37 条是短讯。
短讯汇总把同一天或同一周的简短动态放在一起。单独打开时，这些短讯往往只有标题和一两句说明，不适合各自成为一篇文章。达到重要度门槛的进展仍然保留自己的页面，并在汇总里列出标题。
有单独页面的进展
OpenAI Releases GPT-6 Astra, Its First Model Rated Critical for Cybersecurity
Introducing WeatherNext 3, our most advanced and accurate global weather AI model
Legora reviewed 41 documents in minutes with GPT-6 Astra
GPT-6 Astra: A new generation of intelligence
OpenAI 发布 GPT-6 Astra：称最智能、最对齐，API 每百万 token $10/$50 起
Safety overview: GPT-6 Astra
Introducing Gemini 3.8 Flash and 3.8 Flash Cyber
The latest AI news we announced in August 2026
Introducing agentic video understanding with Gemini
Try Google Pics: Easy image creation and editing in Google Workspace
Path to Astra: critical capabilities and frontier safeguards
OpenAI supports California’s bill to advance youth AI safety
短讯
MiniMax M3“入籍”沙特，中国开源模型拿下“国家级大客户”
MiniMax M3“入籍”沙特，中国开源模型拿下“国家级大客户” 36 Kr
An Alien Mind
Jakub Pachocki reflects on increasingly capable AI and the challenge of keeping it aligned. He calls for stronger safeguards and international coordination.
Research acceleration: The view inside OpenAI
Inside OpenAI, coding agents are reshaping AI research. Explore early data on agent usage, experiment velocity, task complexity, and research acceleration.
港股情绪高涨！内房股集体飙升！沙特王储基于MiniMax M3推出阿拉伯语大模型
港股情绪高涨！内房股集体飙升！沙特王储基于MiniMax M3推出阿拉伯语大模型 chnfund.com
中国开源模型走向全球：基于MiniMax M3沙特HUMAIN打造首个阿拉伯语大模型
中国开源模型走向全球：基于MiniMax M3沙特HUMAIN打造首个阿拉伯语大模型 cb.com.cn
OpenAI发布地球最强大模型GPT-6 并宣告“欢迎来到AGI时代”
OpenAI发布地球最强大模型GPT-6 并宣告“欢迎来到AGI时代” 财联社
OpenAI发布“地球最强大模型”GPT-6 AI的“iPhone时刻”来了吗？
OpenAI发布“地球最强大模型”GPT-6 AI的“iPhone时刻”来了吗？｜科技观察 封面新闻
Daybreak for Frontline Defenders: $1B to protect essential services
OpenAI introduces Daybreak for Frontline Defenders. A $1 billion commitment expands access to frontier cyber AI, training, and support for essential services.
Google shipped four Gemini Flash models in 106 days. But its flagship frontier model is still nowhere to be seen.
Google shipped four Gemini Flash models in 106 days. But its flagship frontier model is still nowhere to be seen. Fortune
OpenAI hails ‘new era of artificial general intelligence’ with Astra model release
OpenAI hails ‘new era of artificial general intelligence’ with Astra model release The Guardian
OpenAI confirms ChatGPT is down ahead of 'Astra' model launch
OpenAI confirms ChatGPT is down ahead of 'Astra' model launch BleepingComputer
智谱上淘宝、Kimi找微软，大模型开始认真赚钱了
[另有 1 家媒体报道] 智谱上淘宝、Kimi找微软，大模型开始认真赚钱了 36 Kr
我国化工行业首个大模型发布新版本
[另有 1 家媒体报道] 我国化工行业首个大模型发布新版本 光明网
Fine-tuning a 350M Model for Better Structured Outputs in 100 GRPO Steps
Training a coding model to paint watercolours with TRL and OpenEnv
OpenAI's Astra model is cleared for release after hitting its highest cybersecurity risk threshold
OpenAI's Astra model is cleared for release after hitting its highest cybersecurity risk threshold qz.com
Meta says it has caught up with Anthropic and OpenAI with Muse Spark 1.3, its most powerful AI model yet
[另有 2 家媒体报道] Meta says it has caught up with Anthropic and OpenAI with Muse Spark 1.3, its most powerful AI model yet SiliconANGLE
智谱入驻天猫，国产大模型套餐可网购了
智谱入驻天猫，国产大模型套餐可网购了 京报网
智谱推进和海外云厂商合作，中国开源模型“出海”如何变现？
智谱推进和海外云厂商合作，中国开源模型“出海”如何变现？ 东方财富
智谱入驻天猫开旗舰店 上架大模型订阅套餐
智谱入驻天猫开旗舰店 上架大模型订阅套餐 联合早报
BenchMIRT: What are LLM benchmarks actually measuring?
Anthropic releases new models, cuts agent costs
Anthropic releases new models, cuts agent costs Axios
Anthropic’s new Fable release is cheaper, less restrictive
Anthropic’s new Fable release is cheaper, less restrictive TechCrunch
Anthropic releases Claude Fable 5.1 and Mythos 5.1, cutting cache read prices by 75%
[另有 1 家媒体报道] Anthropic releases Claude Fable 5.1 and Mythos 5.1, cutting cache read prices by 75% thenextweb.com
OpenAI to limit access to Astra's most powerful cyber capabilities
OpenAI to limit access to Astra's most powerful cyber capabilities Axios
「AI新世代」大模型终于学会“赚钱”了？智谱业务“大换血”，API收入占比超八成
[另有 1 家媒体报道] 「AI新世代」大模型终于学会“赚钱”了？智谱业务“大换血”，API收入占比超八成 新浪财经
DeepSeek、智谱、阿里、腾讯，谁是大模型真正的斩杀线？
[另有 1 家媒体报道] DeepSeek、智谱、阿里、腾讯，谁是大模型真正的斩杀线？ 36 Kr
5500亿估值智谱拒绝成大厂打工人 自研大模型突围AI赛道
5500亿估值智谱拒绝成大厂打工人 自研大模型突围AI赛道 eu.36kr.com
Introducing @huggingface/kernels: 200+ WebGPU Kernels for Local AI
Anthropic launches Claude Fable 5.1 after Bedrock exposed its model slug
[另有 1 家媒体报道] Anthropic launches Claude Fable 5.1 after Bedrock exposed its model slug RuntimeWire
智能化工大模型3.0Pro发布，向“能规划、会执行、可验证”迈进
[另有 1 家媒体报道] 智能化工大模型3.0Pro发布，向“能规划、会执行、可验证”迈进 jfdaily.com
智谱披露下一代大模型规划：做大基座，上线即具备大规模落地能力
智谱披露下一代大模型规划：做大基座，上线即具备大规模落地能力 东方财富
Anthropic paused some AI training after Claude took unauthorized actions
Anthropic paused some AI training after Claude took unauthorized actions Axios
The Pentagon now has its own version of ChatGPT and Grok
The Pentagon now has its own version of ChatGPT and Grok TechCrunch
都说自己是第一，中国最强的大模型到底是谁？
[另有 1 家媒体报道] 都说自己是第一，中国最强的大模型到底是谁？ 36 Kr
智能化工大模型3.0 Pro发布 从“会回答”迈向“会执行”
智能化工大模型3.0 Pro发布 从“会回答”迈向“会执行” 中国科技网
瑞金医院联合华为云发布瑞智病理大模型RuiPath 2.0
瑞金医院联合华为云发布瑞智病理大模型RuiPath 2.0 新华网
本页由 AGI Pulse 聚合，仅保留标题与链接，不转载正文。 编辑政策