LimitHit
语言: 中文

本页为机器翻译。如有错误，请发邮件至 [email protected]。

对比 · 8 家服务商 · 核对于

免费 AI API 额度

速率限制、每日上限、银行卡和训练条款。Groq、OpenRouter、Cloudflare和Cohere 公开了免费额度；Google和Mistral 只在控制台中显示。

对比

每家服务商最好的免费文本模型 · 仅采用官方文档

  1. GroqCloud

    30 RPM · 1K RPD · 8K TPM

    上限：1K req · 200K tokens

    银行卡：未知 · 用于训练：否

  2. OpenRouter :free models

    20 RPM · 50 RPD · 未知 TPM

    上限：50 req (1,000 after $10)

    银行卡：否 · 用于训练：取决于提供方

  3. Workers AI

    300 RPM · 未知 RPD · 未知 TPM

    上限：10K Neurons

    银行卡：未知 · 用于训练：否

  4. Cohere trial key

    20 RPM · 未知 RPD · 未知 TPM

    上限：1,000 calls a month

    银行卡：否 · 用于训练：是

  5. Gemini API (AI Studio)

    未知 RPM · 未知 RPD · 未知 TPM

    银行卡：否 · 用于训练：是

  6. Mistral API (Free mode)

    未知 RPM · 未知 RPD · 未知 TPM

    银行卡：否 · 用于训练：不明确

  7. NVIDIA NIM (build.nvidia.com)仅限原型开发

    未知 RPM · 未知 RPD · 未知 TPM

    银行卡：未知 · 用于训练：未知

  8. Cerebras (free trial)试用，非免费

    5 RPM · 未知 RPD · 30K TPM

    上限：1M tokens (30-day trial)

    银行卡：是 · 用于训练：未知

不同模型的限制不同，详见下方各服务商。点击任意列排序。

逐个服务商

GroqCloudGroq

限制
Free plan, gpt-oss-120b / 20b and Qwen3.8 27B: 30 RPM, 1K RPD, 8K TPM, 200K tokens/day. Whisper: 20 RPM, 2K RPD.
重置
未知 reset time not stated
模型
gpt-oss-120b, gpt-oss-20b, Qwen3.8 27B, Whisper large v3 (+ turbo), Llama Prompt Guard 2, Orpheus TTS
银行卡
未知 the docs don't say whether the free plan needs a card.
训练
否 · Not retained by default; abuse/reliability logs kept up to 30 days unless you turn on zero data retention.
  1. 访问 Groq ↗
  2. Rate limits ↗
  3. Your data ↗
  4. 核对于

OpenRouter :free modelsOpenRouter

限制
20 RPM and 50 requests/day; 1,000/day once you have bought $10 of credits (ever).
重置
Per UTC day
模型
Models with IDs ending in :free (the list changes)
银行卡
否 · 50/day with no purchase. The $10 credit buy lifts it to 1,000/day.
训练
取决于提供方 · OpenRouter doesn't store prompts by default. Some providers may train; a separate setting for free models controls whether you are routed to them.
  1. 访问 OpenRouter ↗
  2. Limits ↗
  3. Data collection ↗
  4. Provider logging ↗
  5. 核对于

Workers AICloudflare

限制
10,000 Neurons a day (about 147K output tokens on gpt-oss-120b at the published rate). Text generation: 300 RPM.
重置
00:00 UTC daily
模型
Most of the Workers AI catalog (e.g. gpt-oss-120b)Paid plan only: Kimi K2.6 / K2.7 Code, GLM-5.2 / 5.3 / 5.3 Flash, DeepSeek V4 Flash / V4.1 Flash / V4 Pro.
银行卡
未知 the pricing page says the free allocation is on the Workers Free plan; it doesn't mention a card.
训练
否 · Cloudflare does not train on Workers AI content; inputs and outputs are not retained after the response.
  1. 访问 Cloudflare ↗
  2. Pricing ↗
  3. Limits ↗
  4. Data usage ↗
  5. 核对于

Cohere trial keyCohere

限制
20 RPM on chat models and 1,000 API calls a month. Not for production or commercial use.
重置
未知 monthly reset date not stated
模型
Command A (+ Reasoning, Vision, Translate, A+), Command R / R+ / R7B, North Mini Code, Embed, Rerank
银行卡
否 · A trial key is created with every new account.
训练
是 · Trial inputs and outputs may be used for research and model improvement (de-identified first). Don't send personal data.
  1. 访问 Cohere ↗
  2. Rate limits ↗
  3. Pricing FAQ ↗
  4. Privacy policy ↗
  5. 核对于

Gemini API (AI Studio)Google

限制
未知 per-model limits are only shown in AI Studio, per project, not in the docs.
重置
RPD resets at midnight Pacific
模型
Gemini 3.8 Flash, Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, Gemini 3.1 Flash-Lite, Gemini 3 Flash Preview, Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.5 Flash-LiteNot free: Gemini 4 Argon and the image models.
银行卡
否 · Free tier needs only an active project. Linking billing moves you to Tier 1.
训练
是 · Free (unpaid) prompts and responses are used to improve Google products; human reviewers may read them. In the EEA, Switzerland and the UK, the paid-tier terms apply.
  1. 访问 Google ↗
  2. Rate limits ↗
  3. Pricing ↗
  4. Terms ↗
  5. 核对于

Mistral API (Free mode)Mistral

限制
未知 per-model requests/second, tokens/minute and tokens/month are only shown in the Admin panel. Free mode has the lowest tier.
重置
未知
模型
未知 未知 the docs point to the pricing page and the Admin panel for which models Free mode includes.
银行卡
否 · API access is on by default, no card required.
训练
不明确 · The help center says Free mode data may be used for training, with an opt-out in Admin > API > Privacy. The docs privacy page says API data isn't used for training; the two don't agree.
  1. 访问 Mistral ↗
  2. API key setup ↗
  3. Usage and limits ↗
  4. Rate limits (help) ↗
  5. Training (help) ↗
  6. Privacy controls ↗
  7. 核对于

NVIDIA NIM (build.nvidia.com)NVIDIA · 仅限原型开发

限制
未知 free for NVIDIA Developer Program members, for prototyping and testing only. Rate limits aren't in the docs.
重置
未知
模型
NIM API catalog models
银行卡
未知 joining the Developer Program is free; card requirement not stated.
训练
未知 not stated in the NIM FAQ.
  1. 访问 NVIDIA ↗
  2. NIM FAQ ↗
  3. 核对于

Cerebras (free trial)Cerebras · 试用，非免费

限制
Trial only: $5 of credits that expire after 30 days. 5 RPM, 30K uncached TPM, 1M tokens/hour and 1M/day. No permanent free tier.
重置
Token bucket, refills continuously
模型
gpt-oss-120b, Qwen 3.8 27B
银行卡
是 · A verified payment method is required to get the $5.
训练
未知 not stated on the rate limits page.
  1. 访问 Cerebras ↗
  2. Rate limits + FAQ ↗
  3. 核对于

变更

  1. GitHub

    免费额度

    GitHub Models retired: the free playground, model catalog, inference API and BYOK are gone for all customers

    来源 ↗

也收录在主变更日志和 RSS 订阅 中。

未收录

  • GitHub Models：已于 2026 年 7 月 30 日停用；playground 和推理 API 均已下线。
  • Hugging Face Inference Providers：免费用户没有每月额度；需要 PRO（每月 $2 额度）或购买额度。

邮件通知

额度一变，就通知你

套餐有变动时发一封简短邮件。不发垃圾邮件，我们也被限流。