OpenAI 发布 GPT-6.1 Terra,价格减半,性能直追旗舰。旨在以更低成本推动 AI Agent 在高频工作场景中的落地
Artificial Analysis 智能指数中,GPT-6.1 Terra(Max)得 48 分,低于 Kimi K3.1 Pro 的 55 分
在 Agents' Last Exam 中等推理强度下比上一代高 15 个百分点,单次任务成本降至约三分之一,并达到 GPT-6 Astra 水平
API 定价为输入每百万 Token 1 美元、缓存输入 0.05 美元、输出 5 美元,较 GPT-5.6 Terra 与 GPT-6.1 Sol 降低 50%
有网友实测 tps 惊人,实测约 450tps,最快可达 600 tps
Kimi 回应额度消耗过快问题:非常抱歉,已经修复并重置月额度 25%
Kimi k3.1 发布时,大量订阅用户发现使用 Kimi 后额度消耗过快,套餐额度过少,向 Kimi 提出质疑
Kimi 回应称他们还调整了套餐的参数,并正在减少不必要的子代理调用和输出长度,以确保减缓配额消耗过快的问题
可是我总感觉这个问题的背后,更大的问题是 Kimi 的套餐配额太少了
你在用 Kimi 套餐吗,关于配额问题你怎么看?
We're sorry for the rate limits and unnecessary quota consumption on the K3.1 Pro subscription plan. These were our issues, and we should have fixed them sooner.
Today, every subscriber on every plan gets a one-time usage reset, restoring 25% of their monthly quota (usage below 25% resets to zero), equal to one week of usage. Your usage bar will reflect this adjustment by end of day PST.
We've also adjusted the plan's parameters and are reducing unnecessary subagent calls and output length, so your quota goes further.
Thank you for your patience while we put this right.
本来今天应该是属于 Kimi 的高光时刻,却因为额度太少大家被频频吐槽
咱就是说,这样下去以后中转掺假都得在 Kimi 里掺 Claude 了😢
Kimi K3.1 Pro 五个新套餐两个老套餐额度最新测算:(仅计算 K3.1 Pro)
Go: 49元,Token 额度:无kimi code 忽略
Plus:99元,月额度:73m,5H:3.48M,5H 换算 0.9 刀
倍率:1x
Pro: 199元,月额度:365m,5H:17.4M,5H 换算 4.3 刀
倍率:2x
Max: 699元,月额度:1169M,5H:55.7M,5H 换算 13.9 刀
倍率:5x
Max: 1399元,月额度:2338m,5H:111.3M,5H 换算 27.8 刀
倍率:10x
老Allegretto:199 元,月额度:703M,5H:34.8M,5H 换算 10 刀
倍率:4x
老Allegro:699 元,月额度:2191M,5H:104.3M,5H 换算 25 刀
实际倍率:10x
卧槽,全体起立!一觉醒来全球震惊瘫坐了!Kimi K3.1 正式发布!
国产大模型正式跻身全球大模型第一梯队!脚踢 Anthropic,拳打 OpenAI!
100 万 token 上下文,MOE,2.8T 总参数 / 1040 亿活跃参数,支持多模态
说实话兄弟们,我看完 Kimi 贴出来的跑分结果我都不敢相信,就连外国友人都说这个成绩难以置信!
Kimi K3.1 Pro 在9项上超过 Claude Opus 5.5,特别强的领域包括:
在 Terminal-Bench 4.0 终端交互任务上,Kimi K3.1 Pro 以 65.7 紧咬第一名的 Opus 5.5(66.4)
在更偏真实软件工程长程推理的 SWE Marathon中,Kimi K3.1 Pro 拿下了第一(51.7),领先 Opus 5.5(46.3)和 Fable 5.1(45.6)
在 Agents' Last Exam 中同样以 43.2 排名第一,超越 Opus 5.5 的 42.6
价格方面,K3.1 Pro 仅为上一代 K3 的 2/3
输入 $2 每百万 token,缓存输入 $0.2 每百万token,输出 $10 每百万token
同时推出高速版,
输入 $4 每百万 token,缓存输入 $0.2 每百万token,输出 $20 每百万token
让我们恭喜 Kimi 跻身全球大模型第一梯队🎉
买了 glm plan,智谱有送 4 次重置,现在 glm 额度根本用不完啊
最近在研究怎么自动化运维自己的工具站,搞了好多自动化任务,其中一个就是自动化外链
每天 0 点跑外链任务,早上 8 点给我汇报进展
目前进展还算顺利,都是专挑免费的搞,不过可以发的外链列表要见底了,准备又要找一批外链了
啊?所以说,今晚让 PHP 程序员睡不着的高危漏洞,是用 GPT 5.6 Sol 发现的?
OpenAI模型在 Patch the Planet 计划中帮助研究人员发现了 PHP 语言的关键安全漏洞
OpenAI领导了一个由维护者、研究人员和安全供应商组成的联盟,在披露前花了两个月准备修复和防护措施,这些漏洞被认为是近期影响最严重的漏洞之一
OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites.
Over the past two months, we led a coalition of maintainers, researchers, and security providers to prepare fixes and protections ahead of disclosure. This duration was necessary because the fix was quite intricate and required a major change to the PHP engine. Every site using PHP can now safely upgrade their version to stay secure, in addition to applying mitigations from various security vendors.
Due to the severity of these flaws and PHP’s widespread use, we consider this bug to be one of the most impactful vulnerabilities in recent memory. It also marks a new phase, where AI can find security bugs with internet-scale impact.
This is an example of the Defender’s Window in action: to our knowledge, frontier AI was able to find this vulnerability ahead of attackers, giving defenders time to prepare.
openai.com/index/coordinated…
Readers added context they thought people might want to know
OpenAI may claim a role in the coordinated response and patching process. But the initial discovery and exploit development, however, occurred several months earlier by an independent researcher.
x.com/lcfr_hax/statu…


















