📋All Platforms全部平台
One unified list. Use the region filters or search to narrow down — click any site link to open it in a new tab.
统一列表呈现全部平台。可用地区筛选或搜索框快速定位，点击任意官网链接将新窗口打开。
Platform Details
平台详情
❓FAQ - Free TokensFAQ - 免费Token常见问题
What are free tokens / 免费Token?什么是免费Token？
Free tokens are free API quotas granted by LLM platforms - either permanent free tiers (e.g. Zhipu GLM-4-Flash, SiliconFlow sub-9B models, Baidu ERNIE-Speed/Lite, iFlytek Spark-Lite, Tencent Hunyuan-Lite, Google Gemini Flash) or one-time new-user bonuses (up to 20M tokens).
免费Token是各大模型平台赠送的免费API额度，分为两类：永久免费层（如智谱GLM-4-Flash、硅基流动9B以下模型、百度ERNIE-Speed/Lite、讯飞Spark-Lite、腾讯混元-Lite、Google Gemini Flash）和新人一次性赠送（最高2000万Token）。
Which platforms give the most free tokens to new users?哪家新用户送的免费Token最多？
Zhipu AI and SiliconFlow each give ≈20M tokens (SiliconFlow adds ≈10M more via onboarding tasks), DeepSeek ≈5M, iFlytek ≈5M, plus 15-30% more from Kimi (¥15 coupon + ≈8M bonus) and Tencent Hunyuan (1M+1M, 1-year validity).
智谱AI和硅基流动各送约2000万Token（硅基流动完成任务再送约1000万），DeepSeek约500万，讯飞星火约500万，Kimi送15元券+约800万，腾讯混元送100万通用+100万嵌入（1年有效）。
Which platforms offer free-forever tokens?哪些平台有永久免费的Token？
Zhipu GLM-4-Flash, SiliconFlow (sub-9B), Baidu ERNIE-Speed/Lite, iFlytek Spark-Lite, Tencent Hunyuan-Lite and Google Gemini Flash are free forever. Local deployment (Ollama, LM Studio) is unlimited and completely free.
智谱GLM-4-Flash、硅基流动9B以下模型、百度ERNIE-Speed/Lite、讯飞Spark-Lite、腾讯混元-Lite、Google Gemini Flash均为永久免费；本地部署（Ollama、LM Studio）则无限免费。
Do free token quotas expire?免费Token额度会过期吗？
Yes - most new-user bonuses expire in 30-90 days (DeepSeek ≈30 days, Alibaba Bailian ≈90 days). Permanent free tiers do not expire but have rate limits (RPM/QPS). Spend the soonest-expiring quota first.
会。多数新人赠送额度30-90天清零（DeepSeek约30天、阿里百炼约90天）；永久免费层不过期但有速率限制（RPM/QPS）。建议优先消耗快过期的额度。
Can free tokens be used in production?免费Token能用于生产环境吗？
Free tiers suit development, testing and learning - they lack SLA, priority queues and support. For production workloads, use paid tiers or aggregators like OpenRouter for stability.
免费层适合开发/测试/学习，缺SLA、优先级排队和技术支持。生产环境建议使用付费层或OpenRouter等聚合平台保证稳定。
How to choose the right free token platform?怎么选择适合自己的免费Token平台？
Coding → DeepSeek; long documents → Kimi (256K); speed → Groq/Cerebras; multimodal → Gemini; Chinese writing → Qwen/Kimi/Doubao; RAG → Alibaba Bailian + SiliconFlow. See the use-case table below.
写代码选DeepSeek；长文本选Kimi（256K）；要速度选Groq/Cerebras；多模态选Gemini；中文写作选通义/Kimi/豆包；RAG选百炼+硅基流动。详见下方场景推荐表。
💡Recommended Combos by Use Case按场景推荐组合
Strategy: use "free-forever" models for daily light calls, harvest one-time new-user bonuses, then move high-frequency / team workloads to aggregators or paid tiers.
组合思路：用「永久免费模型」覆盖日常轻量调用，用「新人高额赠送」一次性薅足，再把高频、稳定、团队化需求切到聚合平台或付费层。
| Use case | 场景 | Recommended setup | 首选方案 |
|---|---|---|---|
| Beginners / testing free tiers | 新手入门 / 白嫖测试 | SiliconFlow (20M tokens) + Zhipu AI (GLM-4-Flash free forever) | 硅基流动（2000 万 Token）＋ 智谱 AI（GLM-4-Flash 永久免费） |
| Coding / AI agents | 写代码 / 跑 Agent | DeepSeek (best for code) + Zhipu GLM-4-Flash as fallback | DeepSeek（编程强）+ 智谱 GLM-4-Flash 兜底 |
| Long documents / analysis | 长文本 / 文档分析 | Kimi (256K context) + Alibaba Bailian | Kimi（256K 上下文）+ 阿里云百炼 |
| Max speed / real-time chat | 极致速度 / 实时对话 | Groq / Cerebras (LPU & wafer-scale inference) | Groq / Cerebras（LPU、晶圆级极速推理） |
| Multimodal (image/audio/video) | 多模态（图文音视频） | Google Gemini 2.5 / ModelScope | Google Gemini 2.5 / 魔搭社区 |
| Chinese writing / content | 中文写作 / 内容创作 | Qwen + Kimi + Doubao | 通义千问 + Kimi + 豆包 |
| Overseas closed-source models | 海外闭源模型 | OpenRouter (unified entry) or OpenRouter + GitHub Models | OpenRouter（统一入口）或 OpenRouter + GitHub Models |
| Knowledge-base RAG | 知识库 RAG | Alibaba Bailian + SiliconFlow | 通义百炼 + 硅基流动 |
| Enterprise / compliance | 企业 / 合规生产 | Baidu Qianfan + Tencent Hunyuan (compliance-first) | 百度千帆 + 腾讯混元（中文合规首选） |
| Data stays local | 数据不出本地 | Self-hosted Ollama with Llama / Qwen (unlimited, private) | Ollama 自部署 Llama / Qwen（无限免费、隐私可控） |