已并入 Buddy AI 产品体系。个人体验版免费，每月含 500 基础积分；WorkBuddy 与 CodeBuddy 使用同一账号并共享积分。

体验版免费，含 2 周专业版试用和 300 Credits；Credits 在 QoderWork CN、Qoder CN CLI、Qoder CN IDE 等全家桶产品间共享。

首次开通后自动发放新人免费额度，有效期 90 天；华北 2（北京）的多款文本与多模态模型通常各含 100 万 Tokens，仅适用于实时推理。具体模型和额度以官方清单及控制台为准。

首次开通后，指定语言模型各提供 100 万 Tokens 免费额度，有效期 3 个月；官网另显示实名新用户可领取 20 元代金券，有效期 1 个月。

Spark Lite 语言模型可免费使用，并提供 OpenAI 兼容的 HTTP API；官方文档另有免费额度领取入口。其他 Spark 版本的免费资源、并发和有效期以产品页及控制台为准。

每位用户每日共 2000 次免费调用，单模型最多 200 次；部分大模型为每日 100 次。注册后需绑定已实名的阿里云账号，模型范围、并发和额度会动态调整。

完成实名认证后可调用平台全部免费模型，免费模型账单消耗为 0，Rate Limits 使用固定值；部分模型同时提供免费原名版本与名称带 `Pro/` 的收费版本。实际免费模型列表和限速以控制台为准。

活动截至 2026 年 12 月 31 日：语言模型和多模态理解模型各提供 100 万 Tokens 免费体验额度，有效期 1 年；图像、视频和 3D 模型另有免费次数或积分。

当前产品页显示：Doubao-Seed-2.1-pro、turbo、Evolving 等文本模型各 50 万 Tokens；视频模型各 200 万 Tokens，图像模型各 200 张，语音、向量和联网资源另有免费额度。

每个火山账号最多创建 1 个免费知识库：免费 30 天，含 100 万 Tokens 模型调用额度，仅对该免费知识库生效，并提供每日 300 次检索。

MiMo Auto 当前限时免费，可匿名、零配置使用终端 AI 编程助手。免费期限、可用模型、速率和后续登录要求可能调整，以小米 MiMo Code 官方项目和产品页面为准。

新用户注册可获得 2000 万 Tokens 和 120 次图像/视频资源包；官网当前另有邀请好友实名注册得 Tokens 活动，最高可累计领取 2 亿 Tokens 资源包。

免费版无需信用卡；新用户注册可一次性获得 20 Credits，自发放起 365 天有效。上传、自动转录、手动剪辑、播放预览和标准导出不消耗 Credits；AI 对话、内容生成和云端渲染按实际操作消耗。

新用户当前可领取 1000 Credits。Miora 是腾讯推出的全场景创意智能体工作室，可在同一画布生成图片、视频、3D、UI/UX 等多模态资产；后续额度以账号页面为准。

新建账户后可获得 5 美元免费 Credits，用于体验 Cerebras 托管模型；官方同时提供免费层限速，当前多款模型最高每日 100 万 Tokens，具体模型和限额可能调整。

每天免费 10,000 Neurons，额度在每日 00:00 UTC 重置；超过后需要升级 Workers Paid，超出免费部分按量计费。

注册后可创建免费的 Evaluation / Trial API Key；多数 Chat 模型限速 20 次/分钟，Trial Key 通常每月最多 1000 次 API 调用。

新用户可获得 1 美元免费 Credits，用于体验 Serverless Inference 等按量计费能力；额度、适用服务和后续计费规则以官方定价页及账户后台为准。

所有 GitHub 账号均包含免费但限速的 GitHub Models 使用量，额度按模型及 Copilot 套餐有所不同，适合原型和实验。

Gemini API 为部分当前模型提供 Free Tier，输入与输出 Tokens 免费，并可免费使用 Google AI Studio；模型、地区和速率限制适用，免费层提交内容可能用于改进 Google 产品。

Free Tier 无需升级付费层；当前 Llama 3.1 8B 示例额度为 30 RPM、每日 14,400 次和每日 50 万 Tokens。不同模型分别限制 RPM、RPD、TPM 和 TPD，准确额度以账户 Limits 页面为准。

免费用户每月获得 0.10 美元 Inference Providers Credits；PRO 用户每月 2 美元，额度可能调整。Credits 仅适用于由 Hugging Face 路由并计费的请求。

Meta 官方页面当前写明：在上线阶段，每个 Meta Model API 账户初始获得 20 美元免费 Credits，可用于测试提示词、比较输出和原型集成；可用模型、地区和后续权益以注册页面为准。

默认启用 Free mode，可创建 API Key 并在 Limits 页面所示限制内使用免费层；适合评估和原型。

NVIDIA Developer Program 会员可免费使用托管 NIM API 端点进行原型开发，并可下载 NIM 用于开发和测试；不适用于生产部署。

累计购买少于 10 Credits 的账户每天可发起 50 次免费模型 API 请求；累计购买至少 10 Credits 后为每天 1000 次。使用 `openrouter/free` 或带 `:free` 后缀的模型，实际可用性、速率和延迟可能变化。

Replicate 提供 Try for Free 模型集合，注册后无需先购买 Credits 即可对指定模型进行有限次数的免费运行；达到免费次数后需要设置计费并购买 Credits，具体次数因模型而异。

Free Tier 当前为部分模型提供 20 RPM、每日 20 次请求和每日 20 万 Tokens；未绑定支付方式的账户会自动使用免费层。模型清单和限额会调整，以官方限额页为准。

新人免费额度通常有效 30～90 天，首次开通新加坡地域后按模型自动发放；仅国际部署范围适用，具体额度以控制台为准。

每个 Vercel 团队账号每月包含 5 美元 AI Gateway 免费 Credits，可用于免费层适用模型；免费额度从首次请求开始计算。购买 Credits 后账号转为付费层，不再享受每月免费额度。

The Free plan requires no credit card. New users receive a one-time grant of 20 credits, valid for 365 days. Uploads, auto-transcription, manual editing, playback and previews, and standard exports do not use credits; AI agent work, content generation, and cloud rendering consume credits per action.

New users can currently claim 1,000 credits. Miora is Tencent’s agentic creative studio for producing images, video, 3D, and UI/UX assets on one canvas; future availability is subject to the account page.

New accounts receive $5 in free credits to try Cerebras-hosted models. Cerebras also publishes free-tier rate limits, currently up to 1 million tokens per day for several models; model availability and exact limits may change.

Includes 10,000 free Neurons per day, resetting at 00:00 UTC. Usage above the allocation requires Workers Paid and is billed by usage.

Registration provides a free evaluation/trial API key. Most Chat models are limited to 20 requests per minute, and trial keys are generally capped at 1,000 API calls per month.

New users receive $1 in free credits to try usage-based services such as Serverless Inference. Credit applicability and subsequent billing terms are subject to the official pricing page and account dashboard.

All GitHub accounts include free, rate-limited GitHub Models usage. Limits vary by model and Copilot plan and are intended for prototyping and experimentation.

The Gemini API offers a Free Tier for selected current models with free input and output tokens plus free Google AI Studio access. Model, region, and rate limits apply, and free-tier content may be used to improve Google products.

The Free tier requires no paid upgrade. A current Llama 3.1 8B example includes 30 RPM, 14,400 requests per day, and 500,000 tokens per day. RPM, RPD, TPM, and TPD limits vary by model; exact limits appear in the console.

Free users receive $0.10 in monthly Inference Providers credits and PRO users receive $2.00, subject to change. Credits apply to requests routed and billed through Hugging Face.

Meta's official developer page currently states that, at launch, every Meta Model API account starts with $20 in free credits for prompt testing, output comparison, and prototyping. Model and regional availability and future terms are subject to the signup experience.

Free mode is enabled by default, allowing API-key creation and free-tier usage within the limits shown in the console; intended for evaluation and prototyping.

NVIDIA Developer Program members receive free access to hosted NIM API endpoints for prototyping and can download NIM for development and testing; this is not a production license.

Accounts with fewer than 10 purchased credits receive 50 free-model API requests per day; purchasing at least 10 credits raises the limit to 1,000 per day. Use `openrouter/free` or a model with the `:free` suffix; availability, rate limits, and latency can change.

Replicate offers a Try for Free collection whose selected models can be run a limited number of times after signup without first purchasing credits. Billing and purchased credit are required after the free-use limit, which varies by model.

The Free Tier currently gives selected models 20 RPM, 20 requests per day, and 200,000 tokens per day. Accounts without a linked payment method use it automatically; the model list and limits may change.

New-user quotas are generally valid for 30–90 days and are granted by model after first activation in the Singapore region. Only International deployment scope is eligible; exact quotas appear in the console.

Each Vercel team account includes $5 per month in free AI Gateway credits for free-tier-eligible models. The allowance starts with the first request; purchasing credits moves the account to the paid tier and ends the monthly free credit.