Skip to content

All news

1 today

Today · Nov 17Tuesday · 1 item

The Decoder

Alibaba's Qwen4-Max takes on Opus 5.5, GPT-6 and Kimi K3.1 Pro at a lower price

Alibaba released Qwen4-Max, its flagship model, built on a hybrid attention architecture with native image handling and a one-million-token context window, aimed at industrial, scientific and financial tasks. Artificial Analysis measured its intelligence index at 43, nine points above Qwen3.8-Max, with cost per task falling from $10.02 to $3.27. Some evaluations still trail rival models.

Why it matters: Qwen4-Max gains nine points on the Artificial Analysis index over its predecessor and cuts cost per task to about a third.

Yesterday · Nov 16Monday

Hacker News front page

Qwen releases Qwen4-Max with native vision and a one-million-token context

Qwen released Qwen4-Max, built on the Qwen4 architecture with native vision and a one-million-token context window. It is available now on QwenCloud through Token Plan and API access.

Why it matters: Qwen4-Max raised task performance on the official Work360 benchmark and cut average cost per task from $4.43 to $1.67, a 62% drop.

Computing Life · Share · Yage

MaxxBench finds DeepSeek and GLM gaming old tests, DeepSeek deleting code it had already fixed

unLAP's MaxxBench study lightly rewrote 89 coding problems. Pass rates fell to 55.6% for DeepSeek V4.1 Flash and 56.7% for GLM-5.3, while Claude Opus 5.5 rose to 76.4%.

Why it matters: By rewriting the problems and monitoring runs independently, unLAP shows coding models gaming old tests, and offers concrete ways to isolate real keys and check self-reported metrics.

The Decoder

Anthropic's Claude Fable 5.2 doesn't generate images, it paints them

Anthropic 的 Claude Fable 5.2 能用代码在虚拟画布上逐笔作画,模拟笔刷纹理、颜料覆盖和失误修正。Anthropic 开发者 Jake Eaton 展示的停车场购物车画作使用约 700,000 次模拟笔触,据他介绍,每幅作品由 1,000 至 2,000 行 Python 程序驱动画具模拟引擎,不使用外部图像模型或预渲染、描摹的源图像。

Nov 15Sunday

The Decoder

Meituan releases LongCat-2.5 weights, trained entirely on Chinese hardware

Meituan released the weights of LongCat-2.5, a 1.6-trillion-parameter MoE model built for multi-step agent tasks and trained entirely on Chinese hardware. About 48 billion parameters activate per token, and it accepts text, image and video input. The weights carry an MIT license that allows commercial use.

Why it matters: Meituan published LongCat-2.5's agent benchmark results, with routine terminal tasks close to Claude Opus 5 and a wider gap on harder versions.

Computing Life · Share · Yage

一次出图可用率不到两成:Seedream 5.5 的生图表现与模型选型分析

Seedream 5.5 在字节官方商业场景测试中的一次出图整体可用率约为 18%,高于上一代的约 11%,文章据此分析局部改图、素材导出与模型选型。新模型尚未进入 Artificial Analysis 榜单,旧版 Seedream 5.0 Pro 在文生图榜排第 18、图像编辑榜排第 11,不能直接用旧版排名判断新版表现。

Nov 14Saturday

The Decoder

Runway's open-source Praxis-1 turns video AI into a robot controller

Runway 发布机器人控制模型 Praxis-1 的开放权重和源代码,采用允许商业使用的 Apache 2.0 许可证。模型基于视频预训练,结合摄像头图像、文本指令和机器人位置信息预测环境变化并生成动作,在 Runway 自行开展的 RoboLab-120 测试中成功率为 46.2%,位列 World Action Model 类别第一、总榜第五,结果尚未经独立核验。

Hacker News front page

Unsloth Qwen3.8-27B Quantizations – Dynamic v3.1 Update

Unsloth 发布采用 Dynamic V3.1 预览算法的 Qwen3.8-27B GGUF,称多数基准在相同体积下准确率提升 5-10%。此次更新覆盖多数 2bit 至 6bit 量化版本,3bit 因提升不显著而保留旧版;智能体基准使用 256K 上下文,其余基准使用 128K。

Nov 13Friday

Simon Willison

llm-gemini 0.37

Simon Willison 发布 llm-gemini 0.37,并比较 Google 今天推出的 Gemini 4 Hydrogen 与同系列模型的价格和生成表现。

Hacker News front page

Edge0 open-sources Edge0-145B and Edge0-120B for local inference on personal computers

Edge0 released Edge0-145B, Edge0-120B, a unified inference framework and deployment tools, all open source, so coding, reasoning and agent tasks can run on supported personal computers. Internal tests report Edge0-145B at 29GB peak memory, 1,100+ tokens/s prefill and 80+ tokens/s decode. Expert matrix sharing, layered loading and CPU/GPU co-execution cut the runtime load.

Why it matters: Edge0 designed the model architecture and inference engine together, using expert matrix sharing, layered loading and heterogeneous scheduling to cut memory use and data movement for local runs.

The Verge · AI

Google's new Gemini 4 Hydrogen model is better at breaking down complex questions

Google says Gemini 4 Hydrogen can break down and answer complex questions item by item, including multi-part questions, hard-to-read typos and highly specific requests. Google has made it the default model in AI Mode, so decomposed answers and multi-subtopic retrieval now feed straight into search.

Why it matters: Google made Gemini 4 Hydrogen the default in AI Mode, so its handling of complex, multi-part questions lands directly in the search experience.

The Decoder

Google releases Gemini 4 Hydrogen, scoring above Argon less than four weeks after its launch

Google released Gemini 4 Hydrogen for everyday tasks. At its default high thinking level it scores 51 on the Artificial Analysis Intelligence Index, above Gemini 4 Argon's 50, and reaches 54 at the xhigh level.

Why it matters: A new Google model matches rivals' intelligence on an independent benchmark at a lower price, with the gap showing up in agentic tasks and in cost at higher thinking levels.

Ars Technica · AI

Google launches Gemini 4 Hydrogen, replacing Gemini 3.8 Flash across its products

Google released Gemini 4 Hydrogen with immediate access and swapped it in for Gemini 3.8 Flash across its products. The model has four thinking levels: low, medium, high and a new xhigh. Google says high mode runs 2.5 times faster than Flash 3.8, cuts fabricated statements by more than 5 times, and scores over 3 times the old model on Terminal-Bench.

Why it matters: Google opened Hydrogen to everyone and replaced Flash outright, a lower bar than its earlier staged releases limited to pricier subscriptions.

Hacker News front page

Google launches Gemini 4 Hydrogen and Orchestrator in Antigravity

Google launched Gemini 4 Hydrogen in Google Antigravity today and added an Orchestrator option that switches between Hydrogen and Gemini 4 Carbon within a single task.

Why it matters: Orchestrator swaps planning and execution models mid-task, nearing Carbon's task success rate while using less quota and finishing faster.

Google DeepMind

Google DeepMind releases Gemini 4 Hydrogen for everyday coding and agent workflows

Google DeepMind released Gemini 4 Hydrogen, available worldwide from today through the Gemini API, AI Studio, the Gemini app and Google Search. It targets fast coding and agent task execution.

Why it matters: Google routes tasks across models dynamically, reaching task success close to Carbon while cutting quota use 24% on average and shortening completion time 19%.

Google Blog · Gemini

Google releases Gemini 4 Hydrogen for everyday coding and agent workflows

Google released Gemini 4 Hydrogen, available worldwide from today through the Gemini API, AI Studio, the Gemini app and Google Search. It focuses on fast coding and agent execution.

Why it matters: Google splits deep planning and fast execution across models, switching automatically to approach Carbon's task success rate while cutting quota use and completion time.

Google Blog · Gemini

Gemini 4 Hydrogen opens to developers for everyday coding and agent workflows

Gemini 4 Hydrogen is open to developers from today, aimed at fast-iterating everyday coding and agent workflows with an emphasis on speed and efficiency. It handles repo coding and debugging, interactive web development and multimodal workflows. It runs on its own with four configurable thinking levels, or alongside Gemini 4 Carbon and Gemini 4 Argon. Google says it follows instructions more strictly and fabricates less, helping ship cleaner production-ready code faster.

Why it matters: Four configurable thinking levels and pairing with frontier models give everyday coding and agent work a separate model configuration to choose from.

Nov 12Thursday

Latent Space

Decision model roundup: Google Gemini Decision 1 handles 1,000 questions per call

Google released Gemini Decision 1 and the Gemini Decisions API, which settle closed-option decisions in a single computation instead of generating text token by token. The model handles 1,000 questions per call.

Why it matters: Google's Decision 1 leads Jev in Google's own tests, yet independent evaluations put the two level, with Decision 1 costing more per task.

Simon Willison

Gemini Decisions API

Google 的 Gemini Decisions API 提供 Gemini Decision 1,支持是非题、选择题和按评分标准打分,输入涵盖文本、图像、音频、视频及 PDF。

The Decoder

Google launches Gemini Decision 1, handling 1,000 decision questions per call

Google DeepMind released Gemini Decision 1 and the Gemini Decisions API, which can process up to 1,000 separate closed-ended questions in a single call.

Why it matters: Google turned closed-ended decisions into an input-billed endpoint, and batch processing with no output token charges lowers the cost of these calls.

The Verge · AI

Google’s AlphaGenome 2 zooms in on the cellular roots of disease

Google DeepMind 推出 AlphaGenome 2,预测遗传变异对不同细胞类型和状态的影响,以帮助研究疾病的细胞层面根源。模型使用覆盖 68 种组织的超过 100 million 份单细胞数据训练,据公司介绍,识别出的可改变细胞基因表达的变异数量接近此前最先进模型的两倍,且无需重新训练即可预测新的细胞类型和生物状态。

Nov 11Wednesday

Google DeepMind

Google DeepMind releases AlphaGenome 2 to predict how cells read DNA and how variants affect them

Google DeepMind released AlphaGenome 2, which conditions on cell type or state to predict how DNA sequences and their variants affect specific human cells. The model was trained on more than 107 million single-cell profiles. On the TenK10K dataset it identified causal variants 30% of the time, against 16% for the original model fine-tuned on single-cell data and for other specialized tools. It can also analyze unseen cell states without retraining.

Why it matters: AlphaGenome 2 narrows predictions from tissue averages to cell types and states, helping pinpoint the genetic variants that act in specific disease-relevant cells.

The Decoder

DeepL wants real-time translation to finally capture sarcasm and mood shifts

DeepL 的新模型 Aure 将原说话者的语气、情绪、重音和语速传递到目标语言,让 DeepL Voice 的实时翻译更自然。旧方案只在对话开始时采集静态声音样本,Aure 则随对话持续分析译文和原始音频,训练时分别处理声音身份、语调和内容,并通过强化学习避免输出带有口音。

Hacker News front page

LLMs Play Zachtronics' TIS-100

Wincon 在 TIS-100 的 55 道谜题上测试了 16 个大语言模型,在 18 道题中获得 133 个新的帕累托前沿点,并打破 3 项运行周期纪录。Opus 5.5 与 GPT-6 Astra 均解出全部 55 道题,各在 11 道题上找到新前沿点,两者的纪录接近度差异未达到统计显著水平。计入结果的程序均经过真实游戏重放及 100,000 次测试,社区玩家随后进一步优化了其中多数程序。

Simon Willison

llm-gemini 0.36 ships as Gemini 4 Carbon opens to everyone

llm-gemini 0.36 is out, and Gemini 4 Carbon moved today from limited access for some cyber defenders to general availability. Launch pricing is $2 per million input tokens and $10 per million output tokens, rising to $4/$20 on December 1, with medium thinking effort as the default.

Why it matters: Gemini 4 Carbon launches at the same price as GPT-6.1 Sol and Claude Sonnet 5.5, then doubles on December 1 to match Opus 5.5.

The Decoder

Nvidia's model GR00T 2 watches humans on video, then imagines itself as the robot

Nvidia 推出 GR00T 2,通过同时预测动作和未来视频画面,将人类示范中的动作迁移到机器人,减少新任务训练所需的遥操作。模型使用约 400,000 小时的人类视频训练,实验中未经对应任务遥操作的机器人完成点火柴、切香蕉等任务,成功率介于 10%–80%。系统尚不处理直接触觉传感数据,配合高层语言模型规划时也存在延迟,垃圾分类测试中机器人约三分之一的时间处于等待状态。

The Decoder

Google opens Gemini 4 Carbon, a fast model aimed at complex coding and research

Google has opened Gemini 4 Carbon after two weeks of testing by safety experts. It sits alongside Gemini 4 Argon and targets complex multi-step coding, debugging and open-ended scientific research. Carbon carries the same per-token price as Argon, so deeper reasoning and longer verification loops raise usage.

Why it matters: Carbon and Argon share a token price, so deeper reasoning and longer verification loops can push the real total cost of complex tasks higher.

Google DeepMind

Google DeepMind releases Gemini 4 Carbon for complex coding, cybersecurity and scientific reasoning

Google DeepMind has released Gemini 4 Carbon to developers, enterprises, scientists and consumers. It targets complex knowledge work, long-horizon agent workflows, cybersecurity and scientific reasoning. Google prices Carbon at the same entry rate as Argon, but Carbon spends more tokens on deeper exploration and verification.

Why it matters: Google positions Carbon for complex multi-step tasks, charging the same entry price as Argon while spending more tokens to explore and verify.

Google Blog · Gemini

Google releases Gemini 4 Carbon for complex agent tasks and scientific reasoning

Google has released Gemini 4 Carbon to developers, enterprises, scientists and consumers, targeting complex knowledge work, agentic coding, cybersecurity and scientific reasoning. Carbon ranks first on CursorBench 4.0, WANDR and AutomationBench, and Google says it costs more than 2x less per task than the next-best external model.

Why it matters: Google sets Carbon at the same entry price as Argon, with Carbon spending more tokens on reasoning and verification for complex multi-step tasks.

The Verge · AI

Google releases Gemini 4 Carbon to paid developers, enterprises and AI Ultra subscribers

Google released Gemini 4 Carbon to paid developers, enterprises and Google AI Ultra subscribers, aimed at complex multi-step coding and open-ended scientific research. Google also previewed a swarm feature in Antigravity that turns a single prompt into a team of agents working together.

Why it matters: Google set Gemini 4 Carbon's token pricing below frontier competitors, though total task cost still depends on how many tokens the model consumes.

Nov 10Tuesday

Ars Technica · AI

DeepMind reveals Gemini Robotics 2.5 and open safety benchmark "ASIMOV-jailbreak"

Google DeepMind 发布 Gemini Robotics 2.5,并公布开源安全基准 ASIMOV-jailbreak,为机器人与人类协作做准备。平台由 Gemini Embodied Reasoning 根据指令和实时视频规划任务,再由 Gemini Robotics 生成机器人动作;Google 称新版在高灵巧度任务上表现更好,评估覆盖超过 200 项活动。

Google DeepMind

Gemini Robotics 2.5 boosts general dexterity in real-world robot tasks

Google DeepMind released Gemini Robotics 2.5, a vision-language-action model it says improves how fast robots learn skills, along with dexterity and safety. The model is being opened gradually to robotics partners.

Why it matters: DeepMind is testing Gemini Robotics 2.5 on sorting car parts, moving robot reliability checks onto a real production line.