{ "schema": "nowverse.reviewed-claims.v1", "reviewedAt": "2026-10-27", "rule": "逐项核对官方正文；仅覆盖声明范围，不证明账号资格、实测性能、安装成功或全站资料均已更新。eventDate 为原文事件日期，未知留空；reviewedAt 是核对日期，不是发布日期。", "subjects": [ { "id": "openai", "title": "OpenAI：模型规格与订阅分账", "docs": [ "vendors/openai", "codex", "codex-desktop" ], "category": "模型与费用" }, { "id": "anthropic", "title": "Claude：当前型号与 API 赠额边界", "docs": [ "vendors/anthropic", "claude-code", "claude-desktop" ], "category": "模型与费用" }, { "id": "deepseek", "title": "DeepSeek：别名、分时价与工具费用", "docs": [ "deepseek", "vendors/deepseek" ], "category": "模型与费用" }, { "id": "google", "title": "Google：Argon 产品入口与引入价", "docs": [ "vendors/google", "antigravity" ], "category": "模型与费用" }, { "id": "glm", "title": "GLM Coding Plan：模型与双周期额度", "docs": [ "glm-coding", "vendors/zhipu" ], "category": "编程与接入" }, { "id": "qwen", "title": "Qwen Code：安装条件与多端入口", "docs": [ "qwen-code" ], "category": "编程与接入" }, { "id": "local", "title": "本地推理不等于整个会话离线", "docs": [ "kb/ollama-vs-lmstudio" ], "category": "本地与安全" }, { "id": "microsoft", "title": "Copilot 与 MXC：推理和执行分开看", "docs": [ "vendors/microsoft" ], "category": "本地与安全" }, { "id": "liquid", "title": "Liquid d1：开放权重决策模型", "docs": [], "category": "历史补档" }, { "id": "openrouter", "title": "OpenRouter：语音入口与已结束促销", "docs": [ "vendors/openrouter" ], "category": "历史补档" } ], "claims": [ { "id": "terra-spec", "subject": "openai", "source": "openai-terra", "eventDate": null, "text": "GPT-6.1 Terra 的 API ID 是 gpt-6.1-terra，支持文本/图像输入、文本输出；上下文 1,050,000 tokens，最大输出 128,000 tokens。工具调用要求 Responses API。", "scope": "官方 API 规格，不代表任意客户端或订阅均可调用。", "evidence": [ { "from": 748, "to": 772, "quote": "Tool calling requires the Responses API." } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "terra-rate", "subject": "openai", "source": "openai-terra", "eventDate": null, "text": "Terra Standard 每百万 token 输入/缓存命中/输出为 $1/$0.05/$5；30 分钟缓存写入为 $1.25。超过 272K 输入时，整次请求输入及缓存费率乘 2、输出乘 1.5；Fast 为 Standard 的 2 倍。", "scope": "美元 API token 单价；不是 ChatGPT Plus 包含用量。Batch/Flex、区域处理和缓存保留时间另有规则。", "evidence": [ { "from": 777, "to": 792, "quote": "Prompts with more than 272K input tokens are priced at 2x input and cache rates and 1.5x output for the full request." } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "codex-billing", "subject": "openai", "source": "codex-pricing", "eventDate": null, "text": "ChatGPT Work 与 Codex 共享订阅用量；API Key 计费独立，不能用 API token 单价推算订阅包含任务数。Plus 为 $20/月，Pro 有 $100/$200/$500 三档。", "scope": "当前官方价格页的美元月价；税费、地区、促销、资格及实际账户额度需另核。", "evidence": [ { "from": 749, "to": 800, "quote": "API token prices are separate from subscription usage; don’t use them to" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "codex-quota", "subject": "openai", "source": "codex-pricing", "eventDate": null, "text": "Pro 当前没有五小时用量限制，但仍可能有周限制。Plus 的 Terra 本地消息估计为每五小时 35–285 条，这只是受任务复杂度、上下文及工具调用影响的估计，并非固定承诺。", "scope": "使用估计，不是保底权益；额度可能调整。", "evidence": [ { "from": 848, "to": 872, "quote": "Pro plans currently have no five-hour limit." } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "codex-retirement", "subject": "openai", "source": "codex-pricing", "eventDate": "2026-10-14", "text": "GPT-5.5 在使用 ChatGPT 登录的 Codex 中于 10/14 19:00 UTC 退役；官方明确 API 与使用 API Key 认证的 Codex 会话不受该次退役影响。", "scope": "限定认证方式的历史事件，不扩大为全部 API 退役。", "evidence": [ { "from": 752, "to": 754, "quote": "key aren’t affected." } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "claude-lineup", "subject": "anthropic", "source": "claude-models", "eventDate": null, "text": "当前模型概览列 Fable 5.1、Opus 5.5、Sonnet 5.5、Haiku 5.5；四款均支持文本/图像输入和文本输出，列出的上下文为 1M、最大输出 128K。", "scope": "官方平台模型概览；其他云平台、客户端和订阅是否开放须分别核对。", "evidence": [ { "from": 36, "to": 49, "quote": "Max output 128K tokens 128K tokens 128K tokens 128K tokens" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "claude-credit-amount", "subject": "anthropic", "source": "claude-credits", "eventDate": null, "text": "Max 5x/20x 的月度 Claude API 赠额分别为 $100/$200。Team 按 Standard 席位 $20、Premium 席位 $100 汇入共享池，每月最多 $500；并非每人 $500。", "scope": "符合条件且有效的计划；新订阅需满 7 天，Free/Pro/Enterprise 不适用。", "evidence": [ { "from": 39, "to": 58, "quote": "Team, Premium seat $100 USD per seat*" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "claude-credit-boundary", "subject": "anthropic", "source": "claude-credits", "eventDate": null, "text": "领取需关联 Claude Console 组织；赠额支持 Claude API、Managed Agents、Agent SDK 和 Playground，不覆盖 Claude Code、Claude 应用额外用量或其他云平台。周期结束过期、不结转；赠额不改变 Claude/Claude Code 使用上限。", "scope": "促销 API 余额与产品订阅用量严格分开；绑定组织及角色条件见官方领取流程。", "evidence": [ { "from": 35, "to": 44, "quote": "Doesn't cover Claude Code, extra usage in the Claude apps" }, { "from": 89, "to": 95, "quote": "Your Claude plan. Credits don't change your usage limits in Claude or Claude Code." } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "deepseek-alias", "subject": "deepseek", "source": "deepseek-pricing", "eventDate": null, "text": "deepseek-flash 对应 DeepSeek-V4.1-Flash-1017；deepseek-v4-pro 对应 DeepSeek-V4-Pro-0813。旧 deepseek-v4-flash 与 deepseek-v4-flash-vision-exp 名称仍可调用，但实际由新版 Flash 服务并按 Flash 计费。", "scope": "别名与实际服务版本不同，不能凭旧名称认定旧模型仍在运行。", "evidence": [ { "from": 33, "to": 57, "quote": "模型版本 DeepSeek-V4.1-Flash-1017 DeepSeek-V4-Pro-0813" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "deepseek-rates", "subject": "deepseek", "source": "deepseek-pricing", "eventDate": null, "text": "每百万 tokens，Flash 空闲时段缓存命中/未命中/输出为 ¥0.02/¥1/¥4，Pro 为 ¥0.15/¥4.5/¥13.5；高峰价均为两倍。高峰为北京时间工作日（不含中国法定节假日）9–12 点与 14–18 点。", "scope": "人民币 API 按量价；不与包月、第三方转售或美元地区价混用。", "evidence": [ { "from": 48, "to": 58, "quote": "高峰时段 8元 27.0元" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "deepseek-search-fees", "subject": "deepseek", "source": "deepseek-pricing", "eventDate": "2026-11-01", "text": "Search 和 Fetch 是独立 API，各按每 1,000 次成功请求 ¥5 计费。通过 Harness/Claude Code/Codex 的服务端工具访问 Search/Fetch，目前未计费；官方声明从北京时间 2026-11-01 00:00 起收费。", "scope": "未来生效安排，不写成 10/27 已经开始收费；模型 token 费用仍独立计算。", "evidence": [ { "from": 60, "to": 68, "quote": "Search 每 1,000 次成功请求 5元" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "argon-access", "subject": "google", "source": "google-argon", "eventDate": "2026-10-13", "text": "10/13 公告分别列出 AI Studio/Gemini platform API、Antigravity 的 Google AI Ultra、Gemini 应用和 AI Mode 的 Pro/Ultra 入口，不能把 Gemini 应用 Pro 权益推广为 Antigravity 同权益。", "scope": "公告描述的产品入口，不证明个人账户、地区或组织已获权。", "evidence": [ { "from": 215, "to": 239, "quote": "for Google AI Ultra subscribers in Google Antigravity" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "argon-intro-price", "subject": "google", "source": "google-argon", "eventDate": "2026-10-13", "text": "Argon 公告列 API 引入价每百万输入 $2、输出 $10，有效至 11/30；随后列为 $4/$20。应同时核对实时计费页，不把引入价写成永久价格。", "scope": "公告里的限期引入价，不是订阅包含额度。", "evidence": [ { "from": 241, "to": 243, "quote": "of $2 per million input tokens and $10 per million output tokens" }, { "from": 309, "to": 309, "quote": "After the introductory period expires on November 30, 2026" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "glm-models", "subject": "glm", "source": "glm-coding", "eventDate": null, "text": "国内 GLM Coding Plan 当前列 GLM-5.3 和 GLM-5.3-Flash。历史 GLM-5.2/5.1 转至 5.3，GLM-5-Turbo/4.7 转至 5.3-Flash；额度仅限官方支持的指定工具环境。", "scope": "bigmodel.cn 国内计划；不自动套用到 z.ai 或任意 API 脚本。", "evidence": [ { "from": 61, "to": 67, "quote": "所有套餐均支持 GLM-5.3 、GLM-5.3-Flash。" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "glm-quota", "subject": "glm", "source": "glm-coding", "eventDate": null, "text": "Lite/Pro/Max 同时受五小时与周积分上限约束，分别为 2,000/10,000、12,000/60,000、28,000/140,000。五小时额度在对应消耗五小时后恢复，周额度自下单起按七天周期刷新。", "scope": "积分不是请求数或固定 token 数；缓存、模型和时段影响消耗。", "evidence": [ { "from": 73, "to": 84, "quote": "Max 套餐 28,000 140,000" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "qwen-install", "subject": "qwen", "source": "qwen-code", "eventDate": null, "text": "Qwen Code 官方 README 的 npm 安装要求 Node.js 22+；命令为 npm install -g @qwen-code/qwen-code@latest。也有独立安装器和 Homebrew 路线，进入 qwen 后通过 /auth 配置 provider 与 API Key。", "scope": "文档核对、未在本机执行安装；下载安装脚本应先检查来源与内容。", "evidence": [ { "from": 50, "to": 82, "quote": "**NPM** (requires [Node.js 22+](https://nodejs.org/)):" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "qwen-interfaces", "subject": "qwen", "source": "qwen-code", "eventDate": null, "text": "Qwen Code 不仅有终端，还列 Desktop、实验性 Web、编辑器与聊天渠道；qwen serve 的 daemon/HTTP+SSE 接入标为实验性。安装可用不代表个人账号免费或模型权限已开通。", "scope": "功能入口清单，不宣称全部接口稳定或已实测。", "evidence": [ { "from": 93, "to": 107, "quote": "**Daemon** — Run `qwen serve`" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "ollama-local-only", "subject": "local", "source": "ollama-faq", "eventDate": null, "text": "Ollama 可设置 OLLAMA_NO_CLOUD=1 或 server.json 的 disable_ollama_cloud=true 后重启，以关闭云模型和网页搜索。默认服务绑定 127.0.0.1:11434；改成对外监听前须另设网络和鉴权边界。", "scope": "关闭 Ollama 云功能不等于其他客户端、工具、下载器也离线；未实测具体模型的硬件性能。", "evidence": [ { "from": 140, "to": 165, "quote": "OLLAMA_NO_CLOUD = 1" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "copilot-local-boundary", "subject": "microsoft", "source": "copilot-local", "eventDate": "2026-10-07", "text": "10/07 Copilot 公告区分 Auto 本地/云端编排与显式本地模型选择；本地推理并不自动使会话离线。Shell、MCP、内置文件工具的隔离边界也不同。", "scope": "公告及渐进开放状态，不宣称所有设备已支持，也不将厂商性能表视为本站实测。", "evidence": [ { "from": 106, "to": 116, "quote": "Model selection, inference, and tool execution have different boundaries; local inference does not make the session offline." } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "mxc-policy", "subject": "microsoft", "source": "microsoft-mxc", "eventDate": "2026-10-07", "text": "MXC 是把文件、网络等声明式策略映射到操作系统隔离边界的执行层，策略位于 Agent 工作负载控制之外。Permissive 只记录原本会拒绝的访问并放行，不能当作已强制隔离。", "scope": "安全设计与模式说明，不是安全认证或对所有攻击免疫的保证。", "evidence": [ { "from": 83, "to": 84, "quote": "The policy remains outside the agent workload’s control" }, { "from": 149, "to": 152, "quote": "In Permissive mode, MXC records access that the policy would have denied but allows the operation to continue." } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "liquid-d1-release", "subject": "liquid", "source": "liquid-d1", "eventDate": "2026-10-07", "text": "Liquid AI 10/07 公布开放权重 d1-3B 与 d1-omni-600M 决策模型；不是通用聊天模型。速度和基准表来自厂商指定硬件与测试条件，不能外推到普通笔记本。", "scope": "开放权重不自动等于任意商用许可；使用前仍核对各模型卡和许可。", "evidence": [ { "from": 16, "to": 28, "quote": "Today, we release d1-3B and d1-omni-600M, two open-weight models" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "openrouter-audio", "subject": "openrouter", "source": "openrouter-elevenlabs", "eventDate": "2026-10-07", "text": "OpenRouter 10/07 公告接入 ElevenLabs 的 9 款 TTS 与 2 款 STT 模型，经 /api/v1/audio/speech 和 /api/v1/audio/transcriptions 调用；无需另买 ElevenLabs 计划，但仍按 OpenRouter 计费。", "scope": "Scribe v2 的实时 WebSocket 不在该次上线范围。", "evidence": [ { "from": 14, "to": 18, "quote": "ElevenLabs is now on OpenRouter with nine Text to Speech models and two Speech to Text models." }, { "from": 39, "to": 41, "quote": "Realtime Scribe v2 over WebSocket isn’t part of this launch." } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" }, { "id": "openrouter-promo-ended", "subject": "openrouter", "source": "openrouter-elevenlabs", "eventDate": "2026-10-19", "text": "该公告的 ElevenLabs 五折促销截止 10/19 08:00 PT，已结束；不能作为 10/27 仍有效的优惠推荐。", "scope": "历史优惠截止时间，不推断后续续期。", "evidence": [ { "from": 18, "to": 18, "quote": "through October 19, 8am PT" } ], "reviewedAt": "2026-10-27", "status": "source-reviewed" } ], "unresolved": [ { "source": "perplexity-embeddings", "state": "access-limited", "note": "官方正文返回 403，不能从搜索摘要或视频转写核准技术与价格声明。", "url": "https://www.perplexity.ai/hub/blog/multimodal-embeddings-beyond-a-single-vector" }, { "source": "minimax-cn-plan", "state": "insufficient-body", "note": "HTTP 200 仅取得站点外壳，未确认本轮套餐正文，不刷新旧价格核对日。", "url": "https://www.minimax.cn/m-plan" }, { "source": "minimax-global-plan", "state": "insufficient-body", "note": "海外计划同样缺实质正文；国内与海外权益不能混用。", "url": "https://www.minimax.io/m-plan" }, { "source": "synthid", "state": "insufficient-body", "note": "仅取得 SynthID Detector 标题，不能确认检测范围、资格或准确率。", "url": "https://synthid.com/" } ], "sources": [ { "id": "openai-terra", "url": "https://developers.openai.com/api/docs/models/gpt-6.1-terra", "finalUrl": "https://developers.openai.com/api/docs/models/gpt-6.1-terra", "checkedAt": "2026-10-27T05:51:39.273Z", "sha256": "327989affa53ea9847ea70e4f7919747bf9d128ccf87c13164fe72370c5a2021", "excerpts": [ { "claimId": "terra-spec", "from": 748, "to": 772, "quote": "Tool calling requires the Responses API.", "text": " GPT-6.1 Terra\n gpt-6.1-terra \n A fast, efficient model for visual reasoning, precise computer use, and clear, natural writing.\n A fast, efficient model for visual reasoning, precise computer use, and clear, natural writing.\n Compare Try in Playground \n Reasoning\n Highest\n Speed\n Fast\n Price\n $1 • $5 \n Input • Output \n Input\n Text, Image\n Output\n Text\n Use GPT-6.1 Terra to match product photos to catalog entries, interpret\ntechnical drawings, act on application screens, or draft and revise text.\n reasoning.effort supports none , low , medium (default), high ,\n xhigh , and max .\n Tool calling requires the Responses API.\n See the GPT-6 guide \nfor prompting and migration guidance.\n 1,050,000 context window\n 128,000 max output tokens" }, { "claimId": "terra-rate", "from": 777, "to": 792, "quote": "Prompts with more than 272K input tokens are priced at 2x input and cache rates and 1.5x output for the full request.", "text": " Text tokens\n Per 1M tokens\n Input\n $1.00\n Cached input\n $0.05\n Cache writes\n $1.25\n Output\n $5.00\n Cached input tokens are priced at 5% of the uncached input token rate.\n Cache-write prices shown use 30-minute retention. Standard cache writes cost $1.25 per million tokens for 30-minute retention and $2.50 for 12-hour retention, before long-context adjustments.\n Prompts with more than 272K input tokens are priced at 2x input and cache rates and 1.5x output for the full request.\n Batch and Flex prices are 50% lower than Standard.\n Regional processing adds a 10% premium where available.\n Fast mode prices are 2x Standard." } ] }, { "id": "codex-pricing", "url": "https://developers.openai.com/codex/pricing", "finalUrl": "https://learn.chatgpt.com/docs/pricing", "checkedAt": "2026-10-27T05:51:39.237Z", "sha256": "6c2a7213111487cf0fc11ae801ee689419a9e971fcbe38b1c96f0af94126356e", "excerpts": [ { "claimId": "codex-billing", "from": 749, "to": 800, "quote": "API token prices are separate from subscription usage; don’t use them to", "text": " ChatGPT Work and Codex are included in your ChatGPT Free, Go, Plus, Pro, Business, Edu, or Enterprise plan\n ChatGPT Work and Codex share usage. ChatGPT Work usage inside\nChatGPT uses the same pricing, credits, and usage limits as Codex.\n GPT-5.5 retires from Codex with ChatGPT sign-in on October 14, 2026, at noon\nPDT (19:00 UTC). The OpenAI API and Codex sessions authenticated with an API\nkey aren’t affected. See GPT-5.5 retirement \nfor migration guidance.\n See token rates for credit-based plans and\n GPT-6.1 Sol model guidance for model details.\nAPI token prices are separate from subscription usage; don’t use them to\nestimate included tasks.\n Pricing options\n Individual Business / Enterprise \n Free\n Explore Codex capabilities on quick coding tasks.\n $0 /month \n Get Free \n GPT-6 Luna at Standard speed in the desktop app, subject to rollout\n Go\n Use Codex for lightweight coding tasks.\n $8 /month \n Get Go \n GPT-6 Luna at Standard speed in the desktop app, subject to rollout\n Plus\n Power a few focused coding sessions each week.\n $20 /month \n Get Plus \n Codex on the web, in the CLI, in the IDE extension, and on iOS\n Cloud-based integrations like automatic code review and Slack\nintegration\n GPT-6.1 Sol and GPT-6 Luna\n Flexibly extend usage with ChatGPT credits \n Other ChatGPT features as part of the\nPlus plan\n Pro\n Choose the Pro plan that fits your usage.\n From $100 /month \n Get Pro \n Everything in Plus and: \n Plans at $100, $200, or $500 USD per month\n Ultrafast mode \naccess on Pro $500\n Other ChatGPT features as part of the\nPro plan\n Compare Pro plans and usage limits. \n API Key\n Great for automation in shared environments like CI.\n Learn more \n Codex in the CLI, SDK, or IDE extension\n No cloud-based features (GitHub code review, Slack, etc.)\n Model availability follows the API models available to your key\n Pay for Codex usage based on API pricing " }, { "claimId": "codex-quota", "from": 848, "to": 872, "quote": "Pro plans currently have no five-hour limit.", "text": " What are the usage limits for my plan? \n The number of messages you can send depends on the model used, size and\ncomplexity of your tasks, and whether you run them locally or in the cloud.\nSmall scripts or routine functions may consume only a fraction of your\nallowance, while larger projects, long-running tasks, or extended sessions that\nrequire the agent to hold more context will use significantly more per message.\n Tasks that look similar can consume different amounts of your allowance. Model\nchoice, context, reasoning, tool use, retrieval, and caching all affect usage,\nso prompt length alone isn’t a reliable estimate.\n For model recommendations, see Models .\n The estimates below show local messages per five-hour period for Plus and\nStandard Business. Pro plans currently have no five-hour limit. Cloud tasks\nmay use more of your allowance than local messages. Usage depends on the model\nand task. These estimates are not fixed message limits; check your\n usage dashboard for current limits\nand reset times.\n Model Plus\n Standard Business\n GPT-6 Astra 5-45 5-45 \n GPT-6.1 Sol 15-160 15-160 \n GPT-6.1 Terra 35-285 35-285 \n GPT-6 Sol 15-150 15-150 \n GPT-6 Luna 350-3,000 350-3,000 \n Local messages and cloud chats share your plan’s usage allowance. Weekly\nlimits may also apply." }, { "claimId": "codex-retirement", "from": 752, "to": 754, "quote": "key aren’t affected.", "text": " GPT-5.5 retires from Codex with ChatGPT sign-in on October 14, 2026, at noon\nPDT (19:00 UTC). The OpenAI API and Codex sessions authenticated with an API\nkey aren’t affected. See GPT-5.5 retirement " } ] }, { "id": "claude-models", "url": "https://platform.claude.com/docs/en/about-claude/models/overview", "finalUrl": "https://platform.claude.com/docs/en/models/overview", "checkedAt": "2026-10-27T05:51:41.510Z", "sha256": "8cbfe26d5d5ebd55a00e96c360bd2ac4ec8e8377dbd14dfba4c885c4d702a328", "excerpts": [ { "claimId": "claude-lineup", "from": 36, "to": 49, "quote": "Max output 128K tokens 128K tokens 128K tokens 128K tokens", "text": " If you're unsure which model to use, start with Claude Opus 5.5 for most workloads. Use Claude Fable 5.1 for demanding reasoning and long-horizon agentic work, or when your evals on Claude Opus 5.5 at higher effort still fall short. All current models support text and image input, text output, multilingual capabilities, vision, and tool use. Each model's page lists the platforms it's available on.\n Feature \n Claude Fable 5.1 For demanding reasoning and long-horizon agentic work \n Claude Opus 5.5 For long-running agentic coding and knowledge work \n Claude Sonnet 5.5 The best combination of speed and intelligence \n Claude Haiku 5.5 For high-volume, latency-sensitive tasks such as classification, extraction, and routing \n Comparative latency Slower Moderate Fast Fastest \n Pricing $10 / input MTok $50 / output MTok $4 / input MTok $20 / output MTok $2 / input MTok $10 / output MTok From $0.10 / input MTok From $0.50 / output MTok \n Claude API ID claude-fable-5-1 claude-opus-5-5 claude-sonnet-5-5 claude-haiku-5-5 \n Capabilities \n Thinking Adaptive (always on) Adaptive (always on) Adaptive Adaptive \n Default effort high medium high medium \n Context window 1M tokens 1M tokens 1M tokens 1M tokens \n Max output 128K tokens 128K tokens 128K tokens 128K tokens " } ] }, { "id": "claude-credits", "url": "https://platform.claude.com/docs/en/about-claude/api-credits-for-subscribers", "finalUrl": "https://platform.claude.com/docs/en/about-claude/api-credits-for-subscribers", "checkedAt": "2026-10-27T05:51:40.979Z", "sha256": "6158c09cce052dd16164d0d8f628747b1589f6a169c0bdbd1d4c6b9a5d4b2f2e", "excerpts": [ { "claimId": "claude-credit-amount", "from": 39, "to": 58, "quote": "Team, Premium seat $100 USD per seat*", "text": " Eligible plans Max 5x, Max 20x, and Team, including discounted Team plans \n Monthly amount $100 USD (Max 5x), $200 USD (Max 20x), up to $500 USD (Team, pooled*) \n Covers Claude API, Claude Managed Agents, the Claude Agent SDK, and playground \n Doesn't cover Claude Code, extra usage in the Claude apps \n Refresh Each billing cycle (or monthly for annual plans) \n Expiry End of each billing cycle (or monthly for annual plans), with no rollover \n Payment method for Claude Platform Not required \n API credit amounts  \n Plan Monthly credit \n Max 5x $100 USD \n Max 20x $200 USD \n Team, Standard seat $20 USD per seat* \n Team, Premium seat $100 USD per seat* \n Discounted Team plans (Nonprofit, Scientists) Same as Team: $20 USD per Standard seat, $100 USD per Premium seat* \n *On Team plans, credits for all seats are pooled into one monthly balance, capped at $500 USD. Each monthly credit, including the first, is based on the seats on your plan at the start of that billing month. For example, a team with three Standard seats and two Premium seats receives $260 USD a month. Adding another Premium seat raises the next month's credit to $360 USD.\n Eligibility  \n Your eligible plan must be active and in good standing.\n New subscribers can claim credits after 7 days on an eligible plan.\n Plans purchased through Claude for iOS or Claude for Android are eligible. Claim your credits on claude.ai in a web browser.\n Free, Pro, and Enterprise plans aren't eligible." }, { "claimId": "claude-credit-boundary", "from": 35, "to": 44, "quote": "Doesn't cover Claude Code, extra usage in the Claude apps", "text": " Claude Max and Team plans include monthly credits for the Claude API. To claim them, you link a Claude Console organization to your plan. Credits then arrive in that organization each billing cycle (or monthly for annual plans), and you can use them to build and run your own applications and agents. You don't need to add a payment method to Claude Platform.\n  Claude Platform on AWS, Amazon Bedrock, Google Cloud, and Microsoft Foundry: These credits apply only to the Claude API in the Claude Console. They can't be used on other platforms.\n Summary  \n Topic Details \n Eligible plans Max 5x, Max 20x, and Team, including discounted Team plans \n Monthly amount $100 USD (Max 5x), $200 USD (Max 20x), up to $500 USD (Team, pooled*) \n Covers Claude API, Claude Managed Agents, the Claude Agent SDK, and playground \n Doesn't cover Claude Code, extra usage in the Claude apps \n Refresh Each billing cycle (or monthly for annual plans) \n Expiry End of each billing cycle (or monthly for annual plans), with no rollover " }, { "claimId": "claude-credit-boundary", "from": 89, "to": 95, "quote": "Your Claude plan. Credits don't change your usage limits in Claude or Claude Code.", "text": " How credits are applied  \n Timing. Credits arrive each billing cycle, shortly after your plan payment is processed. On annual plans, credits arrive monthly.\n Order. Included credits are spent first, then any credits you've purchased. Auto-reload is based only on your purchased balance. It reloads when that balance reaches your threshold, even if included credits remain.\n Expiry. Unused credits expire at the end of each billing cycle (or monthly for annual plans) and don't roll over.\n Sharing. Every API key and workspace in the linked organization draws from the same balance. To cap spending on a project, give it its own workspace and set that workspace's spend limit .\n Visibility. Your credits appear in the Promotional credits section of the Billing page ( platform.claude.com/settings/billing ) in the Console, with the amount and expiry date.\n Your Claude plan. Credits don't change your usage limits in Claude or Claude Code." } ] }, { "id": "deepseek-pricing", "url": "https://api-docs.deepseek.com/zh-cn/quick_start/pricing", "finalUrl": "https://api-docs.deepseek.com/zh-cn/quick_start/pricing/", "checkedAt": "2026-10-27T05:51:42.288Z", "sha256": "556097caba8adfeeed6455714bffd731c7f60db520fcee06a922eba5840663a3", "excerpts": [ { "claimId": "deepseek-alias", "from": 33, "to": 57, "quote": "模型版本 DeepSeek-V4.1-Flash-1017 DeepSeek-V4-Pro-0813", "text": " 模型 deepseek-flash (1) deepseek-v4-pro \n BASE URL (OpenAI 格式) https://api.deepseek.com \n BASE URL (Anthropic 格式) https://api.deepseek.com/anthropic \n 模型版本 DeepSeek-V4.1-Flash-1017 DeepSeek-V4-Pro-0813 \n 思考模式支持非思考与思考模式（默认）\n切换方式详见思考模式 \n 上下文长度 1M \n 输出长度最大 384K \n 功能 Json Output 支持支持 \n Tool Calls 支持支持 \n Responses API 支持支持 \n Anthropic API 支持支持 \n 对话前缀续写（Beta）支持支持 \n FIM 补全（Beta）仅非思考模式支持仅非思考模式支持 \n 图像理解支持不支持 \n 价格 (2) 百万tokens输入\n（缓存命中）空闲时段 0.02元 0.15元 \n 高峰时段 0.04元 0.30元 \n 百万tokens输入\n（缓存未命中）空闲时段 1元 4.5元 \n 高峰时段 2元 9.0元 \n 百万tokens输出空闲时段 4元 13.5元 \n 高峰时段 8元 27.0元 \n 并发限制 (3) 2500 500 \n (1) 模型名请使用 deepseek-flash 。旧模型名 deepseek-v4-flash 、 deepseek-v4-flash-vision-exp 仍可调用，但对应模型已下线，请求将由 DeepSeek-V4.1-Flash-1017 模型提供服务，并按 Flash 价格计费。" }, { "claimId": "deepseek-rates", "from": 48, "to": 58, "quote": "高峰时段 8元 27.0元", "text": " 价格 (2) 百万tokens输入\n（缓存命中）空闲时段 0.02元 0.15元 \n 高峰时段 0.04元 0.30元 \n 百万tokens输入\n（缓存未命中）空闲时段 1元 4.5元 \n 高峰时段 2元 9.0元 \n 百万tokens输出空闲时段 4元 13.5元 \n 高峰时段 8元 27.0元 \n 并发限制 (3) 2500 500 \n (1) 模型名请使用 deepseek-flash 。旧模型名 deepseek-v4-flash 、 deepseek-v4-flash-vision-exp 仍可调用，但对应模型已下线，请求将由 DeepSeek-V4.1-Flash-1017 模型提供服务，并按 Flash 价格计费。\n (2) 空闲时段价格为高峰时段价格的一半。北京时间周一至周五（不含中国法定节假日）9:00 - 12:00、14:00 - 18:00 为高峰时段；其余时段，包括周末及中国法定节假日全天均为空闲时段。" }, { "claimId": "deepseek-search-fees", "from": 60, "to": 68, "quote": "Search 每 1,000 次成功请求 5元", "text": " 网页搜索与抓取价格 \n Search 和 Fetch 是独立的 API，按请求次数计费。\n API 计费单位价格 \n Search 每 1,000 次成功请求 5元 \n Fetch 每 1,000 次成功请求 5元 \n 请注意，当前用户在 DeepSeek Harness/Claude Code/Codex 中使用 DeepSeek API 时，已经会通过 Server 端工具调用访问到 Search/Fetch 接口，这部分 Search/Fetch 接口调用目前未计费，并将于北京时间 2026 年 11 月 1 日 00:00 起按次收取调用费用。\n 接口用法与完整计费规则详见网页搜索与抓取。\n 扣费规则 \n 模型调用费用 = token 消耗量 × 模型单价；Search 和 Fetch 按上表中的成功请求次数计费。对应的费用将直接从充值余额或赠送余额中进行扣减。" } ] }, { "id": "google-argon", "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-argon/", "finalUrl": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-argon/", "checkedAt": "2026-10-27T05:51:42.283Z", "sha256": "8b666d0c2cbc55a9ec4092279251f779a93610d92579876ca1f59f5f1c00914b", "excerpts": [ { "claimId": "argon-access", "from": 215, "to": 239, "quote": "for Google AI Ultra subscribers in Google Antigravity", "text": " Gemini 4 Argon: Built to solve your hardest problems\n Oct 13, 2026\n | \n x.com \n Facebook \n LinkedIn \n Mail \n Copy link \n We’re giving developers, enterprises, scientists, and consumers access to our latest frontier model.\n Koray Kavukcuoglu\n SVP, Google DeepMind and Chief AI Architect, Google\n Share\n x.com \n Facebook \n LinkedIn \n Mail \n Copy link \n In this article \n Built to solve your hardest problems\n Solving complex engineering tasks\n Transforming critical business and professional workflows\n Accelerating real-world impact with researchers and domain experts\n Developing Argon responsibly and safely\n Get started today\n Gemini 4 Argon is now globally available today for developers and enterprises via the API in AI Studio and Gemini platform , for Google AI Ultra subscribers in Google Antigravity , and Google AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search ." }, { "claimId": "argon-intro-price", "from": 241, "to": 243, "quote": "of $2 per million input tokens and $10 per million output tokens", "text": " The model launches today at an introductory price \n 1 \n of $2 per million input tokens and $10 per million output tokens (with cached input tokens at 95% off)." }, { "claimId": "argon-intro-price", "from": 309, "to": 309, "quote": "After the introductory period expires on November 30, 2026", "text": " After the introductory period expires on November 30, 2026, the price of $4 per 1M input tokens and $20 per 1M output tokens will apply. Applies to the Argon model intended for public access only. " } ] }, { "id": "glm-coding", "url": "https://docs.bigmodel.cn/cn/coding-plan/overview", "finalUrl": "https://docs.bigmodel.cn/cn/coding-plan/overview", "checkedAt": "2026-10-27T05:51:43.128Z", "sha256": "b462ba45e7031cdf0bd4a8cb1d1e61030c147863d26fd40f9c40dea5685c4ca1", "excerpts": [ { "claimId": "glm-models", "from": 61, "to": 67, "quote": "所有套餐均支持 GLM-5.3 、GLM-5.3-Flash。", "text": " 所有套餐均支持 GLM-5.3 、GLM-5.3-Flash。\n 调用历史模型 GLM-5.2、GLM-5.1 都将自动切换至 GLM-5.3，调用 GLM-5-Turbo、GLM-4.7 将自动切换至 GLM-5.3-Flash。\n \n 适用工具 \n 套餐仅限在官方支持的指定工具与产品环境中使用。在除规定工具外调用 API，不可享用 Coding 套餐的额度。\n 套餐支持 OpenClaw 使用，但采用次级调度与尽力交付策略，Coding Agent 任务享有资源抢占优先权，高负载下 OpenClaw 任务将自动触发包括动态排队、限流等公平使用策略。\n 订阅套餐后，在上述工具中调用 GLM 模型，按接入指南配置即可使用套餐额度。当套餐额度耗尽后，需要等待下一个 5 小时周期恢复额度，系统不会继续消耗您的其他资源包/账户余额。" }, { "claimId": "glm-quota", "from": 73, "to": 84, "quote": "Max 套餐 28,000 140,000", "text": " 套餐同时设有每 5 小时和每周额度上限，您可以在用量统计中查看您的额度消耗进展： \n 套餐类型 5 小时积分每周积分 \n Lite 套餐 2,000 10,000 \n Pro 套餐 12,000 60,000 \n Max 套餐 28,000 140,000 \n 积分刷新规则 \n 5 小时积分：采用动态刷新机制，积分额度在请求消耗 5 小时后刷新重置。\n 周积分：自套餐下单时起，以 7 天为一个周期刷新。\n 套餐升级与续费：已订阅用户升级套餐后，积分额度将按新套餐档位立即刷新；续费不会触发积分额度重置，原有积分刷新周期保持不变。\n \n 积分抵扣计算方式 \n 模型消耗积分数=（输入 Token × Input 抵扣系数 + 缓存命中 Token × Cached Input 抵扣系数 + 输出 Token × Output 抵扣系数） / 10000" } ] }, { "id": "qwen-code", "url": "https://raw.githubusercontent.com/QwenLM/qwen-code/main/README.md", "finalUrl": "https://raw.githubusercontent.com/QwenLM/qwen-code/main/README.md", "checkedAt": "2026-10-27T05:51:43.242Z", "sha256": "4a0ebab59ff1b820bcc4456fa8c44e1c8593d7fef26a6183c7c18aa36c6b1966", "excerpts": [ { "claimId": "qwen-install", "from": 50, "to": 82, "quote": "**NPM** (requires [Node.js 22+](https://nodejs.org/)):", "text": "

\n

NPM / Homebrew

\n\n**NPM** (requires [Node.js 22+](https://nodejs.org/)):\n\n```bash\nnpm install -g @qwen-code/qwen-code@latest\n```\n\n**Homebrew** (macOS / Linux):\n\n```bash\nbrew install qwen-code\n```\n\n

\n\n## Quick Start\n\nOpen a terminal in your project and start Qwen Code:\n\n```bash\ncd /path/to/your-project\nqwen\n```\n\nInside the session, run `/auth` to configure your provider and API key. Then try:\n\n```text\nExplain this repository and show me where to start.\n```\n\nSee the [Authentication Guide](https://qwenlm.github.io/qwen-code-docs/en/users/configuration/auth/) and [Settings Reference](https://qwenlm.github.io/qwen-code-docs/en/users/configuration/settings/) for detailed setup." }, { "claimId": "qwen-interfaces", "from": 93, "to": 107, "quote": "**Daemon** — Run `qwen serve`", "text": "Choose the interface that fits your workflow:\n\n| Interface | Get started |\n| ------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| **Terminal** | Run `qwen` in your project — see [Quick Start](#quick-start). |\n| **Desktop** | [Download Qwen Code Desktop](https://github.com/QwenLM/qwen-code/releases/tag/desktop-latest) for macOS, Windows, and Linux. |\n| **Web** | Run `qwen serve --open` to open the [Web UI](https://qwenlm.github.io/qwen-code-docs/en/users/qwen-serve/) in your browser. _(experimental)_ |\n| **Editor** | Set up [VS Code](https://qwenlm.github.io/qwen-code-docs/en/users/integration-vscode/), [Zed](https://qwenlm.github.io/qwen-code-docs/en/users/integration-zed/), or [JetBrains](https://qwenlm.github.io/qwen-code-docs/en/users/integration-jetbrains/). |\n| **Chat** | Connect [Telegram](https://qwenlm.github.io/qwen-code-docs/en/users/features/channels/telegram/), [DingTalk](https://qwenlm.github.io/qwen-code-docs/en/users/features/channels/dingtalk/), [WeChat](https://qwenlm.github.io/qwen-code-docs/en/users/features/channels/weixin/), or [Feishu](https://qwenlm.github.io/qwen-code-docs/en/users/features/channels/feishu/), then run `qwen channel start`. |\n\nFor automation and custom integrations:\n\n- **Headless** — Run `qwen -p \"...\"` in scripts, CI/CD, or batch jobs.\n- **SDKs** — Build with [TypeScript](./packages/sdk-typescript/README.md), [Python](./packages/sdk-python/README.md), or [Java](./packages/sdk-java/qwencode/README.md).\n- **Daemon** — Run `qwen serve` to connect clients over HTTP + SSE (ACP). _(experimental)_ [Daemon guide](https://qwenlm.github.io/qwen-code-docs/en/users/qwen-serve/)." } ] }, { "id": "ollama-faq", "url": "https://docs.ollama.com/faq", "finalUrl": "https://docs.ollama.com/faq", "checkedAt": "2026-10-27T05:51:45.694Z", "sha256": "a2d0de62e0e7395356a6839cfc74be0a0feea41257103b76b502c3625f9d804a", "excerpts": [ { "claimId": "ollama-local-only", "from": 140, "to": 165, "quote": "OLLAMA_NO_CLOUD = 1", "text": " Ollama runs locally. We don’t see your prompts or data when you run locally. When using cloud-hosted models, we process your prompts and responses to provide the service but do not store or log that content and never train on it. We collect basic account info and limited usage metadata to provide the service that does not include prompt or response content. We don’t sell your data. You can delete your account anytime. \n \n How do I disable Ollama’s cloud features? \n Ollama can run in local only mode by disabling Ollama’s cloud features. By turning off Ollama’s cloud features, you will lose the ability to use Ollama’s cloud models and web search. \n Set disable_ollama_cloud in ~/.ollama/server.json : \n { \n \"disable_ollama_cloud\" : true \n } \n You can also set the environment variable: \n OLLAMA_NO_CLOUD = 1 \n Restart Ollama after changing configuration. Once disabled, Ollama’s logs will show Ollama cloud disabled: true . \n \n How can I expose Ollama on my network? \n Ollama binds 127.0.0.1 port 11434 by default. Change the bind address with the OLLAMA_HOST environment variable. \n Refer to the section above for how to set environment variables on your platform. \n \n How can I use Ollama with a proxy server? \n Ollama runs an HTTP server and can be exposed using a proxy server such as Nginx. To do so, configure the proxy to forward requests and optionally set required headers (if not exposing Ollama on the network). For example, with Nginx: \n server { \n listen 80 ; \n server_name example.com; # Replace with your domain or IP \n location / { \n proxy_pass http://localhost:11434; \n proxy_set_header Host localhost:11434; \n } \n } " } ] }, { "id": "copilot-local", "url": "https://commandline.microsoft.com/local-models-sandboxed-tools-github-windows/", "finalUrl": "https://commandline.microsoft.com/local-models-sandboxed-tools-github-windows/", "checkedAt": "2026-10-27T05:51:44.748Z", "sha256": "42d3825d2882f519ac454806287d377d49fae7a025a7b8d57202d49e10064ced", "excerpts": [ { "claimId": "copilot-local-boundary", "from": 106, "to": 116, "quote": "Model selection, inference, and tool execution have different boundaries; local inference does not make the session offline.", "text": " Two ways to use local models in GitHub Copilot\n GitHub Copilot is adding two ways to use local models across the GitHub Copilot CLI, Copilot app, and VS Code. Developers can let Copilot’s intelligent Auto orchestration choose when to use local or cloud inference, or they can explicitly select a local model for workflows that require direct control.\n Model selection, inference, and tool execution have different boundaries; local inference does not make the session offline. \n With Auto, developers do not need to decide where each task should be run. Across a multi-turn session, Copilot can consider task context and cache state as it routes work between local and cloud models, preserving useful, cached work as the session evolves. \n This orchestrated experience complements direct model selection, giving developers a choice between letting Copilot optimize model placement and choosing a specific local model themselves. \n Explicit local-model selection supports workflows that need a specific provider, model, or endpoint. Developers can select MAI Code 1.1 Flash through the Windows ML provider or connect GitHub Copilot to OpenAI-compatible local endpoints and choose from the models those endpoints expose.\n How sandboxes help secure tool execution\n An agent’s shell commands normally inherit the access of the account running them. Moving inference onto the device doesn’t change that. Sandboxing applies a policy to the processes and local services the agent launches, controlling access to files, networks, credentials, system capabilities, and execution paths regardless of which model requested the work.\n GitHub Copilot uses  Microsoft Execution Containers , or MXC, an open-source library from the Windows team that translates policy into native operating-system controls. On Windows, GitHub Copilot uses the BaseContainer tier of the ProcessContainer backend. On macOS, it uses Seatbelt. On Linux, it uses  bubblewrap . These local backends don’t require a separate virtual machine or container image, but we plan to make them options available through MXC in the future. \n When sandboxing is enabled, shell commands and, by default, local Model Context Protocol servers and language servers run inside the process boundary. Built-in file tools run inside GitHub Copilot itself: the agent harness checks their requests against the effective policy, but those checks aren’t OS-enforced child-process isolation. Remote MCP servers are also outside the local process sandbox; when MCP sandbox controls apply, GitHub Copilot checks their connection policy in process. \n Opening the GitHub Copilot CLI and running `/sandbox` slash command allows you to configure your settings at any time." } ] }, { "id": "microsoft-mxc", "url": "https://blogs.windows.com/windowsdeveloper/2026/10/07/microsoft-execution-containers-policy-driven-containment-for-ai-agents/", "finalUrl": "https://blogs.windows.com/windowsdeveloper/2026/10/07/microsoft-execution-containers-policy-driven-containment-for-ai-agents/", "checkedAt": "2026-10-27T05:51:44.581Z", "sha256": "7dff5dc706fc1b8f7fcceb76c1b58bdd46050c7032d4617f83139faf567775f6", "excerpts": [ { "claimId": "mxc-policy", "from": 83, "to": 84, "quote": "The policy remains outside the agent workload’s control", "text": " Microsoft Execution Containers (MXC) is a policy-driven execution layer for untrusted code or dynamically generated workloads. In agentic scenarios, developers can use MXC to contain model-generated output, plugins, tools, agent harness, or the entire agent. This limits what the agent workload can access by containing the scope of the impact if something goes wrong. \n Developers declare the resources a workload requires, such as files and network destinations, and MXC enforces the resulting boundary with the appropriate container. The policy remains outside the agent workload’s control, so the agent or generated code cannot grant itself additional access." }, { "claimId": "mxc-policy", "from": 149, "to": 152, "quote": "In Permissive mode, MXC records access that the policy would have denied but allows the operation to continue.", "text": " Observe agent activity without enforcing policy \n In Enforcement mode, MXC applies the policy without activity report. Granted operations proceed, and operations outside the boundary are restricted.\n In Learning mode, MXC continues to enforce the configured boundary. An operation that has not been granted is blocked and recorded in a JSON activity report. This lets developers or IT reproduce containment failures and understand which resources the workload attempted to access.\n In Permissive mode, MXC records access that the policy would have denied but allows the operation to continue. This is useful during policy authoring because the workload can be completed while developers or IT collect evidence about the resources it uses. Permissive mode does not bypass other applicable operating system or organizational restrictions." } ] }, { "id": "liquid-d1", "url": "https://www.liquid.ai/blog/d1-open", "finalUrl": "https://www.liquid.ai/blog/d1-open", "checkedAt": "2026-10-27T05:51:44.819Z", "sha256": "39e52aea5182a38e71a7ca0d0548df01aa12dc528c60283eb7b769fe123b579e", "excerpts": [ { "claimId": "liquid-d1-release", "from": 16, "to": 28, "quote": "Today, we release d1-3B and d1-omni-600M, two open-weight models", "text": " Open d1: Edge decision models for text, vision, and audio\n OCT 7, 2026 \n On this page\n Architecture and Training \n Benchmarks \n Fast Inference Everywhere \n Open d1 in Action \n Get Started \n Citation \n Today, we release d1-3B and d1-omni-600M, two open-weight models in our d1 decision model family . \n d1-3B scores 48.57 on the Decision Index v0.2.1 (public split), ahead of every model under 10B and on par with Decider 35B-A3B, a decision model 12x its size. It runs the full NVIDIA stack, from DGX in the data center to Jetson at the edge: d1-3B answers a question in 8 ms on an NVIDIA GeForce RTX 4090, 16 ms on a Jetson AGX Thor, and 26 ms on a Jetson AGX Orin. Even the Jetson Orin Nano runs it in 50 ms, fast enough for real-time decisions on the smallest edge hardware. \n d1-omni-600M is our first experimental checkpoint, handling both text and image, as well as text and audio. It scores 15.95 on the same index.\n d1-3B and d1-omni-600M models are available today on Hugging Face. Check out our docs on how to run them locally." } ] }, { "id": "openrouter-elevenlabs", "url": "https://openrouter.ai/blog/announcements/elevenlabs-on-openrouter/", "finalUrl": "https://openrouter.ai/blog/announcements/elevenlabs-on-openrouter/", "checkedAt": "2026-10-27T05:51:45.054Z", "sha256": "16fc769ef2369a98211774648a82747413744749d96a4ccd3a1dc540a003b132", "excerpts": [ { "claimId": "openrouter-audio", "from": 14, "to": 18, "quote": "ElevenLabs is now on OpenRouter with nine Text to Speech models and two Speech to Text models.", "text": " Sasha Blumenfeld · 10/7/2026 \n Which model to use Which model to use Walkthrough: from a first MP3 to a voice agent Tips for your first requests Getting started \n ElevenLabs is now on OpenRouter with nine Text to Speech models and two Speech to Text models.\n If your app already talks to a language model through OpenRouter, it can now speak and listen too. Narrate long-form content with direction over tone and delivery with Eleven v4, give your agent faster spoken replies with v4 Turbo, or transcribe conversations with speaker labels with Scribe v2. You don’t need a separate ElevenLabs plan. Just call POST /api/v1/audio/speech and POST /api/v1/audio/transcriptions with your OpenRouter API key.\n Every ElevenLabs model is 50% off OpenRouter’s list price for all OpenRouter customers through October 19, 8am PT " }, { "claimId": "openrouter-audio", "from": 39, "to": 41, "quote": "Realtime Scribe v2 over WebSocket isn’t part of this launch.", "text": " Scribe v2 elevenlabs/scribe-v2 General transcription with word timestamps, speaker labels, and audio-event tags in 90+ languages \n Scribe v2 Medical elevenlabs/scribe-v2-medical Medical terms with 35% fewer errors on clinical audio compared to Scribe v2 \n Both Scribe models use the same request shape and the same per-second price. Realtime Scribe v2 over WebSocket isn’t part of this launch." }, { "claimId": "openrouter-promo-ended", "from": 18, "to": 18, "quote": "through October 19, 8am PT", "text": " Every ElevenLabs model is 50% off OpenRouter’s list price for all OpenRouter customers through October 19, 8am PT " } ] } ], "coverage": { "collected": 42, "reviewedSources": 13, "claims": 23, "note": "仅统计本批限定范围核对；不代表全站或所有采集来源已经审核。" } }