Topic
Frontier models
Generational updates of flagship models, their limits and price-performance.
Follow this topic
On this site · ElsewhereTopic graph
16 nodes · 52 links- Entries in this topic
- Neighbours outside the topic
- Builds on · Successor · Extends
- Prerequisite · Uses
- Explains · Supports
- Related
Skills and knowledge entries are written in Chinese, as are signals from Chinese sources.
Published signals
28 signalsGemini 4 Argon: Built to solve your hardest problems
We’re giving developers, enterprises, scientists, and consumers access to Gemini 4 Argon, our latest frontier model.
MiMo-V2.6 技术报告：小米花 260 万美元跑一次强化学习
小米 MiMo 团队 10 月 8 日发布 MiMo-V2.6 技术报告：Pro（1.02T 总参数、激活 42B）与 Flash（310B、激活 15B）两款 MoE 模型，强化学习同时扩大批量、任务环境与评分算力。每步约 2.5 万条轨迹、27 亿至 37 亿 token，Pro 的强化学习花费约 260 万美元，其中评分器占 12.7%。训练全程冻结 MoE 路由，作弊轨迹控制在 2% 以下；训练动态、强化学习环境与框架开源。
DeepSeek-V4 的周期性盲点：信息换个位置，检索准确率差 40 个百分点
字节跳动 Seed 发现，采用分块 KV 缓存压缩的模型，检索准确率会随信息相对压缩窗口边界的位置周期性起伏。128K 上下文里，DeepSeek-V4-Flash-Base 在不同位置之间最多差 40.2 个百分点，后训练过的 V4.1-Flash 仍差 6.1 个百分点。起伏的周期正好等于压缩步长：V4 是 4 个 token，V4.1 是 2 个。平均分会把这种弱点藏起来。
One Model Family, Two Gold-Level Results: Fine-Tuning Nemotron for IOI and IMO
One Model Family, Two Gold-Level Results: Fine-Tuning Nemotron for IOI and IMO
Claude Haiku 5.5
As previously promised, here's Anthropic's new fast, low cost model: Introducing Claude Haiku 5.5. The previous Haiku, 4.5, was very much showing its age. It came out almost a year ago, and was priced at $1/million input and $5/million output - relatively exp…
Sharing AI progress in mathematics
OpenAI publishes new results on open problems in mathematics from an internal frontier model and shares Lean proof formalizations and research details on GitHub.
Gemini 4 Argon：DeepSWE 等多项领先，先只给网络安全团队用
Google 10 月 1 日发布 Gemini 4 Argon，在 DeepSWE（77.9%）、AutomationBench、法律与金融智能体评测上领先 GPT-6 Astra 与 Opus 5.5，输出上限从 6.4 万 token 提到 100 万。但它在 Terminal-bench 4.0 上只有 57.4%，低于 Opus 5.5 的 66.4% 和 Sonnet 5.5 的 70.6%，在 FrontierSWE 和 Terminal-Bench Science 上也落后于 Astra。首发价每百万 token 输入 2 美元、输出 10 美元，之后涨到 4 / 20 美元；眼下只通过 Fairwind 计划开放给网络安全防御团队，普通开发者还用不上。
Introducing GPT-6.1 Sol
Meet GPT-6.1 Sol: near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra’s standard API input and output token prices.
Claude Sonnet 5.5：终端任务从 10% 到 71%，按任务算未必更便宜
Anthropic 9 月 28 日发布 Claude Sonnet 5.5，每百万 token 输入 2 美元、输出 10 美元，是 Opus 5.5 的一半，也与 GPT-6 Sol 同价。它在 Terminal-Bench 4.0 上从 Sonnet 5 的 10.3% 升到 70.6%，超过 Opus 5.5 的 66.4%，但在 FrontierCode 与 OSWorld 上仍略低于 Opus 5.5。Anthropic 称单任务成本最多低 30%；而 Artificial Analysis 在最高推理档测得单任务 7.60 美元，比 Sonnet 5 高约 50%。
Introducing Gemini 3.8 Live with Live Avatar
Introducing Gemini 3.8 Live with Live Avatar
Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war
Yesterday was Grok 4.7 (pelicans) and MiMo v2.6 Flash/Pro (…
Some thoughts on the Navier–Stokes Millennium Prize Problem
OpenAI's claimed resolution of the Navier–Stokes problem, the priority dispute with Tristan Buckmaster and Levent Alpöge, and the unanswered question of whether their product usage helped train the model.
GPT-6 Astra: A new generation of intelligence
Introducing GPT-6 Astra, our most intelligent and aligned model yet, with state-of-the-art capabilities across computer use, coding, cybersecurity, and science.
菲尔兹奖得主：模型强在找到那个对象，弱在知道何时放弃
1998 年菲尔兹奖得主 Timothy Gowers 把近期最轰动的几项模型数学成果摊在一起看，发现它们几乎全是「构造出一个对象」而不是「证明一条定理」。他随后否掉了最顺口的解释，给出的替代解释适用范围远超数学：模型的两项确定优势是知道得多、失败得起，于是它在「大量试、试错便宜」的路线上占优；而人类暂时保住的那一项，是走到半路能闻出这条路没戏，从而砍掉搜索树的绝大部分。
Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things
Hands-on evaluation of Qwen 3.8 27B on a 128GB M5 Max and an NVIDIA DGX Spark: the shipped xhigh reasoning default burns tokens on trivial prompts, vision bounding boxes are strong, and throughput of 15–30 tokens per second is the real catch.
Qwen3.8-27B 开源：消费级显卡跑得动，尺子却是自己命名的
270 亿参数、原生多模态、262K 上下文可外推到 100 万，Apache 2.0，量化后一张 24GB 消费级显卡装得下——这一代 Qwen 把「本地跑得动的智能体」这条线又往前推了一格。榜单确实好看，但领先幅度最大的那一项跑在一把以厂商自己命名的基准上，而所有对比都由厂商自己完成。
668 阶哈达玛矩阵被构造出来：三个人加 Claude，清掉 2000 阶以下所有未决阶数
Anthropic 研究员、数学家 Levent Alpöge 报告：由三名人类与 Claude 组成的团队构造出 668 阶哈达玛矩阵——2005 年以来最小的未知阶数，人类卡了三十年。他一次放出 12 张矩阵，清掉 2000 阶以下所有悬而未决的阶数。EpochAI 已在 FrontierMath 上暂标为「由 AI 解决」。
Introducing Gemini 3.7 Flash
Gemini 3.7 Flash arrives three weeks after 3.6 Flash with across-the-board gains in coding, web development and document-heavy knowledge work, at an introductory price of half the previous cost per million tokens.
Meta 重回开放权重：一封长文、一个独立董事会，和一组本地部署的数字
Muse Glimmer 以 Apache 2.0 开放权重的同时，Meta 宣布重回开放路线：扎克伯格发表长文反对能力集中，公司设立独立董事会逐项审核每次模型发布，并披露了让本地部署成立的量化与推测解码数字。Muse Spark 1.2 的权重据称随后开放。
Meta is back with Muse Glimmer: local, agentic, multimodal, and open source
Meta released Muse Glimmer, a 30B multimodal model distilled from Muse under Apache 2.0, designed for local agentic use, with day-0 support across transformers, llama.cpp and vLLM.
BigBang-v1：把数据生产系统本身当成优化对象，以及一处对不上的口径
上海交大人工智能学院、深势科技与上海算法创新研究院组成的无尽前沿团队发布 BigBang-v1：35B 参数、推理时约 3B 激活、262K 上下文，后训练数据称 100% 由 AI 自主合成，Apache-2.0 开源。可复用的不是跑分，而是它对任务的两个硬条件——必须同时前沿且可验证；缺一，合成数据都会迅速贬值。附一条核对：报道称它是「首个 RSI 原生训练的基座模型」，而 Hugging Face 上的元数据把它标为 Qwen3.6-35B-A3B 的微调。
Gemini Robotics 2 brings whole body intelligence to robots
Gemini Robotics 2 is a family of three models — a vision-language-action model for whole-body humanoid control, an embodied reasoning model (ER 2) that orchestrates tools and tracks task progress, and an on-device VLA that adapts to a new robot body in hours.
Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber
We're introducing new Gemini models, including Gemini 3.6 Flash, 3.5 Flash-Lite and 3.5 Flash Cyber. Flash Cyber is a lightweight cybersecurity model to find and patch vulnerabilities.
Safety and alignment in an era of long-horizon models
OpenAI shares lessons from deploying long-running AI models, highlighting new safety risks, observed failures, and improved safeguards through iterative deployment.
Kimi K3, and what we can still learn from the pelican benchmark
Moonshot AI announces Kimi K3, a 2.8-trillion-parameter flagship with an open-weight release promised by July 27 — the first 'open 3T-class' model. Third-party evaluation (Artificial Analysis) puts it behind only Claude Fable 5 on long-horizon knowledge work, at a cost per task comparable to GPT-5.6 Sol.
Welcome Inkling by Thinking Machines
Thinking Machines releases Inkling, an open-weight ~1T-parameter (41B active) Mixture-of-Experts multimodal model with a 1M context window, trained on 45T tokens of text, images, audio and video, with day-one support across Transformers, vLLM, SGLang, llama.cpp and Unsloth.
GPT-Red: Unlocking Self-Improvement for Robustness
OpenAI introduces GPT-Red, an automated red-teaming system that uses self-play to continuously probe and harden models against safety, alignment and prompt-injection failures.
GPT-5.6: Frontier intelligence that scales with your ambition
OpenAI's newest frontier model promises more intelligence from every token, stronger performance per dollar, and more capability on demand.