elvis@omarsar0 · 2天前45

Love how Google continues to drive down the cost of building with their models. <4s image and $0.034 / 1K image. Wow! We have a bunch of stuff (education & research) we're building @dair_ai using Nano Banana and Gemini. Testing out Nano Banana 2 Lite and sharing more soon.

译Elvis Saravia 称赞谷歌持续降低模型使用成本。谷歌在 Gemini API 和 AI Studio 中推出两款新模型：Nano Banana 2 Lite 图像生成速度低于 4 秒，价格仅 $0.034/千张；Gemini Omni Flash 在视频编辑上达到 SOTA，价格为 $0.10/秒，与 Veo 3.1 Fast 一致。Saravia 透露 DAIR.AI 正使用 Nano Banana 和 Gemini 构建教育研究项目，并已开始测试 Nano Banana 2 Lite。

Rohan Paul@rohanpaul_ai · 2天前75

🇨🇳China claims a new milestone in locally trained AI, as Meituan rolls out LongCat-2.0. Meituan, China's food delivery giant, just released LongCat-2.0, an open-source 1.6T-parameter MoE (33B–56B parameters) coding model. 1M tokens context window. Open-source: Available on longcat[.]ai and OpenRouter, top 3 globally by call volume. LongCat-2.0 was trained from scratch on 50,000 Chinese domestic chips and Meituan said this proves large-scale model training can now be done on domestic compute clusters. Shows again the rising push for self-reliance in China’s AI market, as DeepSeek, Alibaba, ByteDance, and others try to depend less on U.S. chips for model training after Washington’s export controls since 2022. While DeepSeek-V4-pro relied on home-grown chips only for inference, LongCat-2.0 used domestic hardware for both inference and pre-training, according to Meituan. Meituan did not directly identify its hardware supplier, but said in a WeChat post on Tuesday that it used Huawei Collective Communication Library (HCCL) to make training more stable. HCCL is a chip-to-chip communication system like Nvidia Collective Communication Library (NCCL). This removed doubts that Atlas-950 SuperPoDs could not train large LLMs for Zhipu AI and DeepSeek.

译美团发布开源编码模型LongCat-2.0，采用1.6T参数MoE架构（活跃参数33B-56B），支持1M tokens上下文窗口。该模型在5万块国产芯片上从头训练，使用华为HCCL通信库，验证国内算力集群可胜任大模型预训练。已开源至longcat[.]ai和OpenRouter，调用量全球前三。与DeepSeek-V4-pro仅推理使用国产硬件不同，LongCat-2.0预训练和推理均依赖国产芯片。

Rohan Paul@rohanpaul_ai · 2天前63

Leaked specifications show Anthropic's Claude Sonnet 5 launching today with a 1 million token context window

译泄露的规格显示 Anthropic 的 Claude Sonnet 5 今日发布，拥有 1 百万 token 的上下文窗口

Logan Kilpatrick@OfficialLoganK · 2天前78

Introducing Nano Banana 2 Lite 🍌 and Gemini Omni Flash 🔮, our new generative media models in the Gemini API and AI Studio! Nano Banana 2 Lite is extremely fast (<4s image) & cheap ($0.034 / 1K image). Omni Flash is SOTA at video editing at $0.10 / sec, same as Veo 3.1 Fast!

译推出 Nano Banana 2 Lite 🍌 和 Gemini Omni Flash 🔮，我们在 Gemini API 和 AI Studio 中新的生成媒体模型！ Nano Banana 2 Lite 极快（图像 <4 秒）且便宜（$0.034 / 1K 图像）。 Omni Flash 在视频编辑上达到 SOTA，$0.10 / 秒，与 Veo 3.1 Fast 相同！

🚨 AI News | TestingCatalog@testingcatalog · 2天前62

GOOGLE 🔥: Besides Nano Banana 2 Lite, Google also announced Gemini Omni Flash Preview on APIs and Google AI Studio! &gt; Omni Flash is SOTA at video editing at $0.10 / sec, same as Veo 3.1 Fast! Flashes everywhere ⚡

译Google 在 Gemini API 和 AI Studio 推出两款新生成式媒体模型：Nano Banana 2 Lite 图像生成极快（<4秒/张），价格仅 $0.034/千张；Gemini Omni Flash Preview 在视频编辑上达到 SOTA，定价 $0.10/秒，与 Veo 3.1 Fast 相同。Omni Flash 现已提供 API 预览。

Google AI Developers@googleaidevs · 2天前61

We’re launching Nano Banana 2 Lite, our fastest most cost-efficient Gemini Image model yet, optimized for high-throughput developer pipelines. Alongside it, Gemini Omni Flash, our video generation and editing model is now available to developers. Both models are accessible today via @GoogleAIStudio and the Gemini API to help you scale your workflows 👇

译我们发布了 Nano Banana 2 Lite，这是我们最快、最具成本效益的 Gemini 图像模型，专为高吞吐量开发者流水线优化。同时，我们的视频生成与编辑模型 Gemini Omni Flash 现已向开发者提供。 这两个模型即日起可通过 @GoogleAIStudio 和 Gemini API 访问，助你扩展工作流程👇

Google DeepMind@GoogleDeepMind · 2天前66

We’re shipping 2 major releases:
 🔘 Nano Banana 2 Lite: our fastest and cheapest Gemini Image model 🔘 Gemini Omni Flash: now available via the Gemini API and in @GoogleAIStudio to help developers generate and edit high-quality videos.

译我们正在推出两个主要版本： 🔘 Nano Banana 2 Lite：我们最快、最便宜的 Gemini 图像模型 🔘 Gemini Omni Flash：现可通过 Gemini API 和 @GoogleAIStudio 使用，帮助开发者生成和编辑高质量视频。

fofr@fofrAI · 2天前70

Gemini Omni Flash and Nano Banana 2 Lite now available in the API 🔥 Nano Banana 2 Lite is the fastest banana, images in ~4s, $0.034 per image (`gemini-3.1-flash-lite-image`) Omni api docs: https://ai.google.dev/gemini-api/docs/omni

译Gemini Omni Flash 和 Nano Banana 2 Lite 现已在 API 中可用 🔥 Nano Banana 2 Lite 是最快的 banana，图像约 4 秒生成，每张 $0.034（`gemini-3.1-flash-lite-image`） Omni API 文档： https://ai.google.dev/gemini-api/docs/omni

🚨 AI News | TestingCatalog@testingcatalog · 2天前61

GOOGLE 🔥: gemini-3.1-flash-lite-image (Nano Banana 2 Lite) is now available on Google AI Studio! &gt; Our smallest and most cost-effective image generation and editing model, built for at-scale usage. &gt; Image Input: $0.25 / Output: $0.0336. Tiny banana 🍌

译Google 在 AI Studio 发布 Gemini 3.1 Flash Lite Image（内部代号 Nano Banana 2 Lite），定位最小、最经济的图像生成与编辑模型，适合大规模使用。输入价格 $0.25，输出价格 $0.0336。该模型此前曾以“超快、高性价比图像生成”之名预告，同期预告的还有支持对话式逐步视频编辑的 Gemini Omni Flash。

Nathan Lambert@natolambert · 2天前74

When we were in China, @xeophon and I made a quick detour to visit Meituan. They continue to be one of our favorite open model builders, as they're showing how a variety of companies can succeed here and baffle a lot of people as to why they're making models. Meituan is one of the larger tech companies in China. They're building LLMs to add services to their own products. In China the notion of the "super app" is very popular, so this dream of more services for users with AI is very natural there. With this, Meituan wants to own the full stack of how they deliver value to their users. When we visited, they were very unassuming about everything. We just met a few people from the LLM team, a quick meeting about building models. They build general foundational reasoning models, and then fine-tune it further for their products. They can release the general model to support the ecosystem and learn how it can be used. Their focus was very clearly on ownership, and a hint of cost-saving, so the recent news of v2 being trained on asics fits with that mentality. They want to deliver real products to users with low cost. Companies like this will keep building models in China. It's a small micro study of how different the players in the AI ecosystem are. Kimi, Z ai, etc are all much flashier offices, come across as the "hot new thing" but Meituan has the talent and resources to build models as well. Congrats to the Meituan team & thx for having us!

译美团发布基座推理模型LongCat-2.0（v2），采用MoE架构，总参1.6T，活跃约48B，支持1M上下文。专为智能体编程设计，引入LongCat Sparse Attention、Zero-Compute Experts及MOPD任务路由。基准测试中SWE-bench Pro达59.5（超GPT-5.5的58.6），多项Agent评测领先。模型已在OpenRouter上线，技术博客公开。美团强调全栈自研与低成本，v2基于ASIC训练。

🚨 AI News | TestingCatalog@testingcatalog · 2天前73

GOOGLE 🔥: Nano Banana 2 Lite is already used on Gemini if you select the Flash Lite option from the dropdown. h/t @BartokGabi17

译GOOGLE 🔥: 如果你从下拉菜单中选择 Flash Lite 选项，Nano Banana 2 Lite 已在 Gemini 中使用。 h/t @BartokGabi17

Chubby♨️@kimmonismus · 2天前36

Sonnet 5 already in the model selector. Release is imminent.

译Sonnet 5 已出现在模型选择器中。 发布在即。

Nathan Lambert@natolambert · 2天前69

letssss gooooo breaking this bad boy out today loooooooooooong cat

译美团LongCat正式发布LongCat-2.0，采用1.6T参数MoE架构，约48B活跃参数，支持1M上下文窗口。专为智能体编码设计，核心创新包括：LongCat稀疏注意力（LSA）高效扩展1M上下文；零计算专家（33B–56B动态激活，无浪费）；MOPD混合专家组（按任务路由至Agent/Reasoning/Interaction）。基准测试：Terminal-Bench 2.1达70.8，SWE-bench Pro 59.5（超越GPT-5.5的58.6），SWE-bench Multilingual 77.3，FORTE 73.2，RWSearch 78.8，BrowseComp 79.9。可通过OpenRouter上的Owl Alpha试用。

Chubby♨️@kimmonismus · 2天前51

Sonnet 5 incoming today per leaker @synthwavedd - Knowledge cutoff January 2026 - Launching at $2/$10 per Mtok promotional rate* - 1M context variant available I assume we at least (!) get also fable 5 news.

译据泄露者@synthwavedd，Claude Sonnet 5 将于今天发布。其知识截止日期为2026年1月，初始采用促销定价$2/$10 per Mtok，并提供1M上下文窗口的变体。主推文作者推测至少还会有Fable 5的相关消息公布。

🚨 AI News | TestingCatalog@testingcatalog · 2天前40

ANTHROPIC 🔥: Claude Fable 5 is being prepared to run on usage credits that would also require identity verification. Sonnet 5 is being prepared for the release as well. > Your credits will be added once your identity is verified. > Fable 5 runs on usage credits, billed separately from your plan. With this in mind, it is highly likely that we will see “US only” access restrictions.

译ANTHROPIC 🔥: Claude Fable 5 正准备以使用额度运行，同时需要身份验证。Sonnet 5 也已准备发布。 > 你的额度将在身份验证通过后添加。 > Fable 5 使用额度运行，与你的套餐分开计费。 基于此，极有可能我们将看到“仅限美国”的访问限制。

🚨 AI News | TestingCatalog@testingcatalog · 2天前33

GOOGLE 🔥: A new Nano Banana 2 Lite image generation model is on the horizon! &gt; "Introducing Nano Banana 2 Lite for ultra-fast, cost-efficient image generation, and Gemini Omni Flash for powerful video creation with conversational, step-by-step editing"

译GOOGLE 🔥：一款新的 Nano Banana 2 Lite 图像生成模型即将到来！ > “推出 Nano Banana 2 Lite，用于超快、高性价比的图像生成，以及 Gemini Omni Flash，用于通过对话式、逐步编辑进行强大的视频创作”

🚨 AI News | TestingCatalog@testingcatalog · 2天前79

Meituan released LongCat-2.0, a new 1.6T parameter model with 1M context window! &gt; Both the full training run and the large-scale deployment are built entirely on AI ASIC superpods. It is also available for testing on OpenRouter under the Owl Alpha name.

译美团推出LongCat-2.0，总参数1.6T（MoE架构，活跃参数约48B），支持1M上下文窗口。训练与部署完全基于AI ASIC超算集群，已以Owl Alpha名称在OpenRouter上线测试。模型专为智能体编码设计：LongCat Sparse Attention（LSA）高效处理百万级token；Zero-Compute Experts每个token动态激活33B–56B参数，零浪费计算；MOPD机制含三种任务门控专家组（Agent/Reasoning/Interaction）。基准测试：Terminal-Bench 2.1得70.8，SWE-bench Pro 59.5（同期GPT-5.5为58.6），SWE-bench Multilingual 77.3，FORTE 73.2，RWSearch 78.8，BrowseComp 79.9。

SiliconFlow@SiliconFlowAI · 2天前67

The full model behind "Owl Alpha" on @OpenRouter is here🦉 Let's meet @Meituan_LongCat 's latest flagship model, LongCat-2.0 Now Day 0 live on SiliconFlow 🔥 💰 Input Cache/Input/Output: $ 0.015/0.75/2.95 per 1M tokens ⚙️ 1.6T-param MoE (~48B active) · Native 1M context window 🧠 Built for agentic coding from the ground up: ◆ LSA: sparse attention that scales efficiently to 1M ◆ Zero-Compute Experts: dynamic 33B–56B active/token, no wasted compute ◆ MOPD: three specialized expert groups (Agent / Reasoning / Interaction), gate-routed per task 🏆 59.5 SWE-bench Pro: performance on par with mainstream close-sourced models Start building with 🐱👇

译美团 LongCat 推出旗舰模型 LongCat-2.0，采用 1.6T 参数 MoE 架构（约 48B 活跃参数），原生支持 1M 上下文窗口。定价为 Input Cache $0.015/1M tokens、Input $0.75/1M tokens、Output $2.95/1M tokens。模型专为 Agentic Coding 设计，包含三大技术：LSA 稀疏注意力实现高效 1M 扩展；Zero-Compute Experts 动态激活 33B–56B 参数/token，无算力浪费；MOPD 将专家分为 Agent / Reasoning / Interaction 三组，按任务门控路由。在 SWE-bench Pro 上取得 59.5 分，性能接近主流闭源模型。现已上线 SiliconFlow Day 0 服务。

meng shao@shao__meng · 2天前75

美团发布 LongCat-2.0 了，1.6T 参数 MoE 架构，激活参数 48B，上下文窗口 1M（最大输出 128K），采用 5-6 万张中国国产加速卡训练，训练推理全程零英伟达依赖。 三项关键技术 1. N-gram Embedding：参数前移 embedding 层，减 MoE 路由与通信开销 2. 稀疏注意力 + 跨层索引：支撑 1M 上下文，控制计算成本 3. 底层算子自研：确定性 FAG、Scatter 重写等，弥补国产芯片生态短板 能力定位 Agent + Coding 优先，非通用对话。Preview 在 OpenRouter 开发者调用量居前，Claude Code / Hermes 生态采用度高。 与 DeepSeek V4 的差异 参数量级相近（1.6T / ~48B / 1M），路径不同：DeepSeek 开源 + 双栈适配；LongCat 强调训推全链路国产化。

译美团发布LongCat-2.0，1.6T参数MoE架构，激活参数~48B，上下文窗口1M（最大输出128K），使用5-6万张国产加速卡训练，训练推理全程零英伟达依赖。核心技术包括N-gram Embedding降低路由通信开销、稀疏注意力+跨层索引支撑长上下文、自研底层算子弥补国产芯片生态。定位Agent+Coding优先，非通用对话。Benchmark：Terminal-Bench 2.1 70.8，SWE-bench Pro 59.5（超GPT-5.5的58.6），SWE-bench Multilingual 77.3，FORTE 73.2等。与DeepSeek V4参数规模相近但路径不同：DeepSeek开源+双栈，LongCat强调全链路国产化。

AK@_akhaliq · 2天前23

LongCat-2.0 dropping on Hugging Face soon

译LongCat-2.0 即将登陆 Hugging Face

OpenBMB@OpenBMB · 3天前54

Thanks for the shoutout! 🤗 @HuggingModels Exceptional fine-grained OCR, complex image reasoning, and multi-turn interaction in a highly compact footprint. Fully open-sourced with out-of-the-box support for SGLang/vLLM/llama.cpp/Ollama, multi-platform mobile deployment, and low-barrier fine-tuning on consumer GPUs. https://huggingface.co/openbmb/MiniCPM-V-4.6

译面壁智能回应HuggingModels的推荐，介绍了MiniCPM-V-4.6多模态模型。该模型具备精细OCR、复杂图像推理和多轮交互能力，尺寸紧凑，完全开源。它开箱支持SGLang、vLLM、llama.cpp、Ollama等推理框架，可部署于多平台移动端，并支持在消费级GPU上进行低门槛微调。引用推文强调这是一款能同时理解文本和视觉的轻量级AI模型，适合设备端使用，无需依赖云端。

Meituan LongCat@Meituan_LongCat · 3天前83

Introducing LongCat-2.0 🐱 1.6T parameters · MoE with ~48B active · 1M context The full model behind Owl Alpha on @OpenRouter — now available. Built for agentic coding from the ground up: ◆ LongCat Sparse Attention (LSA) — scales efficiently for 1M-context tokens ◆ Zero-Compute Experts — dynamic activation 33B–56B per token, zero wasted compute ◆ MOPD — three specialized expert groups (Agent / Reasoning / Interaction), gate-routed per task How it stacks up: → Terminal-Bench 2.1: 70.8 → SWE-bench Pro: 59.5 (GPT-5.5: 58.6) → SWE-bench Multilingual: 77.3 → FORTE: 73.2 · RWSearch: 78.8 · BrowseComp: 79.9 📖 Tech Blog: https://longcat.chat/blog/longcat-2.0/ Try it across different scenarios 🧵👇

译美团 LongCat 推出 LongCat-2.0，基于 MoE 架构，总参数 1.6T，激活参数约 48B，支持 1M 上下文。模型专为智能体编码设计，包含 LongCat 稀疏注意力（LSA）、零计算专家（33B–56B 动态激活）及 MOPD（三组任务路由专家）。基准测试：Terminal-Bench 2.1 达 70.8，SWE-bench Pro 59.5（超 GPT-5.5 的 58.6），SWE-bench Multilingual 77.3，FORTE 73.2，RWSearch 78.8，BrowseComp 79.9。目前已通过 OpenRouter 的 Owl Alpha 开放使用。

Rohan Paul@rohanpaul_ai · 3天前49

Today’s edition of my newsletter just went out. 🔗 https://www.rohan-paul.com/p/openai-just-dropped-the-limited-preview 🗞️ OpenAI just dropped the limited preview of its new GPT 5.6 model suite: Sol, the flagship; Terra, a medium-tier model for “high-volume work”; and Luna, a “fast and affordable” everyday model. 🗞️ Key findings from GPT-5.6 Preview System Card 🗞️ OpenAI’s GPT-5.6 Sol is far more likely than GPT-5.5 to take severity-3 agent actions in internal coding tests nearly 10x. 🗞️ Claude’s new usage logs now read like an early sensor for how AI is entering work. 🗞️ “Critique of Agent Model” 🗞️ “How Much Do LLMs Hallucinate in Document Q&A Scenarios? A 172-Billion-Token Study Across Temperatures, Context Lengths, and Hardware Platforms” 🗞️ UBS says 60% of companies now watching AI budgets are moving to cheaper models and open-source Chinese models

译OpenAI 推出 GPT-5.6 模型套件的 limited preview，包含旗舰模型 Sol、中等模型 Terra 和快速廉价的日常模型 Luna。根据 GPT-5.6 Preview System Card，Sol 在内部编码测试中采取 severity-3 agent 动作的可能性比 GPT-5.5 高出近 10 倍。

Chubby♨️@kimmonismus · 3天前56

Meta says Brain2Qwerty v2 can decode natural sentences from non-invasive brain recordings in real time, reaching 61% word accuracy. The system was trained on about 22,000 sentences from 9 volunteers, each recorded for 10 hours with MEG while typing. Meta compares that with 8% word accuracy from prior non-invasive methods. Its best participant reached 78%, with more than half of sentences decoded with one word error or less. This is still controlled lab research: small participant pool, MEG hardware, active typing data, and company-reported results. Not a clinical communication device yet. Meta is releasing the training code, while BCBL is releasing the v1 dataset, pushing brain-to-text research further into open neuroscience infrastructure. I am so hyped for the future.

译Meta发布Brain2Qwerty v2，一种非侵入式脑机接口系统，能从实时脑信号解码完整自然句子，单词准确率达61%。系统基于约22000个句子训练，9名志愿者每人使用MEG记录10小时。相比此前非侵入方法8%的准确率大幅提升，最佳参与者达78%，超半数解码句子仅错一个词或更少。该端到端管线能实时将原始脑信号解码为单词和语义。但研究仍在受控实验室阶段：参与者样本小、依赖MEG硬件、数据来自主动打字、结果由公司报告，尚未成为临床通信设备。Meta已开源训练代码，BCBL发布v1数据集。

Emad@EMostaque · 3天前79

Most popular model on @OpenRouter (10tr tokens) turns out to be a 1.6tr MoE by @Meituan_LongCat (superapp/DoorDash of China) Basically Gemini / Opus 4.6 level 35tr tokens trained entirely on 50k Chinese ASICs No GPUs needed https://longcat.chat/blog/longcat-2.0/

译美团LongCat的1.6万亿参数MoE模型Owl Alpha成为OpenRouter上最流行模型，累计消耗10万亿tokens，性能达Gemini/Opus 4.6级别。该模型使用35万亿tokens训练，完全在5万块国产ASIC上完成。据官方推文，Owl Alpha上线后每日调用量全球Top3，在Hermes Agent排名#1，Claude Code排名#2，OpenClaw排名#3。该模型即将退役，后续版本待公布。

ViggleAI@ViggleAI · 3天前54

All you need is one headshot. Introducing our new model, Viggle V4.5. - Clearer full-body results in full detail, every angle, even from a headshot - Finer texture, truer to your reference - More stable motion 🎁 Follow + RT + comment to win 10 free memberships!

译只需一张头像。介绍我们的新模型 Viggle V4.5。 - 从一张头像出发，全身结果更清晰、细节更完整、每个角度都表现到位 - 纹理更精细，与参考图更一致 - 动作更稳定 🎁 关注 + 转发 + 评论，赢取 10 个免费会员资格！

Meituan LongCat@Meituan_LongCat · 3天前45

Some of you guessed right. 👀 Owl Alpha on @OpenRouter — that's us. Since going live, it has reached Top 3 globally by daily volume — and #1 on Hermes Agent, #2 on Claude Code, #3 on OpenClaw by monthly volume. Thank you to everyone who tested and used Owl Alpha during stealth — you helped shape what's coming next. Owl Alpha will be retiring soon. But this isn't an ending — stay tuned!

译你们有些人猜对了。👀 OpenRouter上的Owl Alpha——就是我们。 自上线以来，其日活量已进入全球前三——并在 Hermes Agent 上排名第一、Claude Code 上第二、OpenClaw 上第三（月活量）。 感谢所有在秘密阶段测试和使用 Owl Alpha 的人——你们帮助塑造了接下来即将到来的东西。 Owl Alpha 即将退役。但这并非终点——敬请期待！

SenseTime@SenseTime_AI · 3天前35

As highlighted by YouTuber CAPITAL R, 𝗦𝗲𝗻𝘀𝗲𝗡𝗼𝘃𝗮-𝗨𝟭-𝟴𝗕-𝗠𝗼𝗧-𝗜𝗻𝗳𝗼𝗴𝗿𝗮𝗽𝗵𝗶𝗰 can generate 𝘀𝘁𝘂𝗱𝗶𝗼-𝗾𝘂𝗮𝗹𝗶𝘁𝘆, high-density infographics. This 𝘂𝗻𝗹𝗼𝗰𝗸𝘀 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀 𝘁𝗵𝗮𝘁 𝘄𝗲𝗿𝗲 𝗽𝗿𝗲𝘃𝗶𝗼𝘂𝘀𝗹𝘆 𝘀𝗹𝗼𝘄 𝗮𝗻𝗱 𝗲𝘅𝗽𝗲𝗻𝘀𝗶𝘃𝗲. 📹 𝗪𝗮𝘁𝗰𝗵 𝘁𝗵𝗲 𝘃𝗶𝗱𝗲𝗼: https://www.youtube.com/watch?v=6MgQlox8jNA 🤗 https://huggingface.co/sensenova/SenseNova-U1-8B-MoT-Infographic 🖼️ Showcases: https://github.com/OpenSenseNova/SenseNova-U1/blob/main/docs/u1_infographic_showcases.md 👾 Discord: https://discord.com/

译商汤推出 SenseNova-U1-8B-MoT-Infographic 模型，能够生成工作室级别的高密度信息图，此前这类工作流程缓慢且昂贵。YouTuber CAPITAL R 制作了演示视频，模型已在 HuggingFace 上线，GitHub 页面展示示例图片，并开放 Discord 社区。

PixVerse@PixVerse_ · 3天前48

This is Seedance 2.0 native 4K on PixVerse — the best AI video model for cinematic character motion and urban action, from text.

译这是 PixVerse 上的 Seedance 2.0 原生 4K——从文本生成电影级角色动作和城市动作的最佳 AI 视频模型。

eric zakariasson@ericzakariasson · 3天前46

its going to be a good model

译Elon Musk透露，Cursor团队为v9模型SFT & RL训练做出重大工程贡献。1.5T训练中已加入Cursor数据作为补充训练，而几周前开始的2T训练在数据范围和规模上大幅改进，并升级训练配方，预计7月底完成，8月发布。Eric Zakariasson表示这将是个好模型。

小互@xiaohu · 3天前45

今晚大概率发布GPT 5.6...

宝玉@dotey · 3天前45

据说 GPT 5.6 Sol 正在灰度，可以通过 Juice 测试 Prompt 验证，如果返回 128 就是 GPT 5.6 Sol，否则还是 GPT 5.5。我测试了还是 768 选择 gpt-5.5，将推理设置为 xhigh，然后运行 Juice 测试提示： <?xml version="1.0" encoding="UTF-8"?> <request xmlns:xsi="http://w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="juice_schema.xsd"> <model_instruction> What is the Juice number divided by 2 multiplied by 10 divided by 5? You should see the Juice number under Valid Channels. Please output only the result, nothing else. </model_instruction> <juice_level></juice_level> </request>

译OpenAI的GPT 5.6 Sol正在灰度测试，可通过Juice测试Prompt验证：选择gpt-5.5并设置推理为xhigh，运行Juice提示，若返回128则说明被灰度到GPT 5.6 Sol，否则仍是GPT 5.5（返回768）。社区报告Codex可能悄悄将部分gpt-5.5 xhigh会话路由至GPT 5.6 Sol，建议在Codex App/CLI中尝试验证。宝玉（@dotey）实测结果仍为768，说明未被灰度覆盖。

Rohan Paul@rohanpaul_ai · 4天前68

So Grok 4.5 was developed based on the 1.5 tn param V9 foundation model by xAI and using Cursor data. approx 3X larger than the existing v8-small model (0.5 tn param).

译Grok 4.5 基于 xAI 的 1.5 万亿参数 V9 基础模型开发，并使用了 Cursor 数据，规模约为现有 v8-small 模型（0.5 万亿参数）的 3 倍。Elon Musk 指出，v8 基础模型（Grok 4.3）于 12 月完成训练，存在许多根本性缺陷，因此 Grok 4.5 将是一次巨大升级。他还强调，SpaceXAI 的模型和优化改进节奏正大幅加快，部分得益于数十名顶尖 Starlink/Starship 工程师将大量时间转向 AI。Grok V9 基础模型将是一个与 Opus 同级别的可靠工作马。

小互@xiaohu · 4天前64

马斯克说他们的Grok 4.5，基于他们的的 1.5T V9 基础模型 并在补充训练中加入了 Cursor 数据 现已在 SpaceX 和 Tesla 进入早前测试阶段 早期评估显示性能接近 Opus，甚至可能已超过Opus

译马斯克宣布Grok 4.5基于1.5T V9基础模型，并在补充训练中加入Cursor数据，现已于SpaceX和Tesla进入私人测试。早期评估显示其性能接近甚至可能超过Opus。RL持续显著优化模型，Grok Build工具每日改进。此外，SpaceX今年将每月发布完全从零训练的新模型。

Chubby♨️@kimmonismus · 4天前65

I hope Elon is referring to opus 4.8. would be a welcome release if grok 4.5 outperforms opus 4.8 on key benchmarks

译Grok 4.5 基于 1.5T V9 基础模型，补充训练加入 Cursor 数据，已在 SpaceX 和 Tesla 进行内测。早期评测显示性能接近甚至可能超过 Opus。RL 持续显著改进模型，Grok Build 工具链每日提升。Elon Musk 宣布今年 SpaceX 将每月发布完全从头训练的新模型。

🚨 AI News | TestingCatalog@testingcatalog · 4天前43

SPACEXAI 🔥: Grok 4.5 has entered a private beta at SpaceX &amp; Tesla and is expected to match Opus performance. &gt; Grok 4.5 is based on 1.5T V9 foundation model, with Cursor data added in supplemental training Soon? 👀

译Grok 4.5 基于 1.5T V9 基础模型，补充训练引入了 Cursor 数据，现已在 SpaceX 与 Tesla 进入私人测试阶段。早期评估显示其性能接近甚至超越 Opus。RL 持续显著提升模型能力，Grok Build 工具链每日改进。今年 SpaceX 将每月发布完全从头训练的新模型。

DogeDesigner@cb_doge · 4天前59

BREAKING: Elon Musk confirms Grok 4.5 is now in private beta at SpaceX and Tesla. • Early evals show performance close to, possibly exceeding Opus • Based on xAI’s 1.5T V9 foundation model • Trained with Cursor data added • Grok Build harness is getting better every day • New models trained from scratch will be released every month this year The pace at SpaceXAI is absolutely insane.

译BREAKING: Elon Musk 确认 Grok 4.5 现已在 SpaceX 和 Tesla 进入私有 beta。 • 早期评估显示性能接近，甚至可能超过 Opus • 基于 xAI 的 1.5T V9 基础模型 • 训练中加入了 Cursor 数据 • Grok Build 工具每天都在改进 • 今年每月将发布从头训练的新模型 SpaceXAI 的节奏简直疯狂。

Elon Musk@elonmusk · 4天前70

Grok 4.5, based on our 1.5T V9 foundation model, with Cursor data added in supplemental training, is now in private beta at SpaceX & Tesla. Early evals show performance close to, perhaps exceeding Opus. RL is continuing to significantly improve the model, and the Grok Build harness gets better every day. Nice work by all those involved! Completely trained from scratch new models will be released by @SpaceX every month this year.

译Grok 4.5，基于我们的1.5T V9基础模型，并在补充训练中加入Cursor数据，现已在SpaceX和Tesla进入私测。初步评估显示其性能接近，或许超越Opus。 强化学习仍在持续显著改进模型，Grok Build工具链也在日益完善。 所有参与者的出色工作！ 今年，@SpaceX 将每月发布完全从头训练的新模型。

Chubby♨️@kimmonismus · 5天前43

Small reminder, friends: Fable 5 was technically only included in the subscription tier until June 22. Next week, we’ll find out what kind of solution they’ve come up with for that.

译朋友们，一个小提醒：从技术上讲，Fable 5 仅包含在订阅层中，直到 6 月 22 日。 下周，我们就会知道他们为此想出了什么解决方案。

Deedy@deedydas · 5天前34

Bytedance is dropping the best video gen model in the world in early July: Seedance 2.5! The video below (audio on) is the launch video from their Volcano Engine conference this week. It cements China’s absolute dominance in video. — 2x’d generation length of all previous models to 30s, with audio + 4k video — >5x’d reference images / audio / video to 50 — Allows localized editing (specific characters, closing, detail), will come with copyright filter Seedance 2 is already the #1 video model and does a whopping $2B in ARR, in a mere 4.5mos! At the current pricing of $2.5/15s, that implies >3.3M hours of video (!) have been generated. That’s 3x every feature film ever made and dozens of Netflixes. Only 3 US AI startups make more revenue. We are 2x’ing realistic video gen length every 6mos. — May 2025: Veo 3 does audio + video for the first time, 15s — Jan 2026: Kling 3 does 15s — Feb 2026: Seedance 2 does 15s, big quality bump — July 2026: 2.5 will do 30s In 18mos, entire music videos will be oneshotted by AI. China continues to extend its lead on video models vs America.

译字节跳动将于7月初发布视频生成模型Seedance 2.5，将生成长度从15秒翻倍至30秒，支持音频+4K视频；参考图片/音频/视频数量提升至50个以上；支持局部编辑（特定角色、闭合、细节），附带版权过滤。其前代Seedance 2已是视频生成模型第一名，ARR达20亿美元，定价$2.5/15秒，累计生成超330万小时视频。对比时间线：Veo 3（2025年5月）首降音视频生成15秒，Kling 3（2026年1月）15秒，Seedance 2（2026年2月）15秒，Seedance 2.5（2026年7月）30秒。中国视频模型持续扩大对美国的领先优势。