Pricing
AI API pricing links
Review each provider's billing model, top-up notes, and pricing page. Exact model rates should be checked on the provider site, docs, or dashboard.
19
Provider pricing links
| Provider | Type | Models | Billing | Top-up | Checked | Link |
|---|---|---|---|---|---|---|
| 云雾 API | China relay | Claude / OpenAI / Gemini / DeepSeek | The pricing page lists model prices; exact model rates and discounts should be checked on the pricing page or dashboard. | Top-up and payment are handled in the platform dashboard. | 2026-06-25 | Open |
| API 易 | China relay | Claude / OpenAI / Gemini / DeepSeek / Qwen | Overseas models follow official pricing, domestic models are lower than official rates, and rare models may include a small markup. | Uses a fixed 1:7 exchange rate with a $5 / RMB 35 minimum top-up and top-up bonus tiers. | 2026-06-25 | Open |
| OpenRouter | Global router | Claude / OpenAI / Gemini / DeepSeek / Llama | Pay-as-you-go access to 400+ models; the pricing page states there is no minimum spend or lock-in. | Credits are purchased and managed in OpenRouter; taxes may apply by region. | 2026-06-25 | Open |
| DMXAPI | Global router | GPT / Claude / Llama / Image models | Backend billing follows official provider prices, while recharge is offered at a discount. | Discounted recharge; exact top-up tiers should be checked on DMXAPI. | 2026-06-25 | Open |
| V-API | China relay | OpenAI / Gemini / Claude / DeepSeek | The pricing page lists prices for ChatGPT, DeepSeek, Gemini, Claude, and related models. | Top-up and payment are handled in the V-API dashboard. | 2026-06-25 | Open |
| 柏拉图 AI | China relay | OpenAI / Claude / Midjourney / Suno / Luma | The platform covers OpenAI, Claude, Midjourney, Suno, Luma, and related models; model multipliers may change with platform notices. | The platform provides a top-up entry; tiers and payment methods should be checked in the dashboard. | 2026-06-25 | Open |
| New API | Self-hosted | OpenAI / Claude / Gemini / DeepSeek / Qwen | Self-hosted gateway pricing depends on upstream channels, group multipliers, and model multipliers. | Supports internal quota and top-up configuration, including EPay and Stripe integrations. | 2026-06-25 | Open |
| One API | Self-hosted | OpenAI / Claude / Gemini / DeepSeek / Qwen | Self-hosted pricing is configured by the deployer through channels and user groups. | Supports quota management, redemption codes, recharge links, and initial user balance settings. | 2026-06-25 | Open |
| AI API 中转站导航 | China relay | OpenAI / Claude / Gemini / DeepSeek / Qwen | Directory site with categories and pricing comparison links; it does not provide a unified billing plan. | Top-up depends on the selected provider. | 2026-06-25 | Open |
| PackyCode | China relay | Claude / GPT / Gemini / Qwen / GLM | PackyAPI is usage-based on the gateway side, while the public site also presents a separate Codex monthly plan. | RMB top-up; balance bonuses, discounts, and monthly-plan pricing should be checked on PackyCode. | 2026-06-25 | Open |
| Requesty | Global router | OpenAI / Anthropic / Google / Meta | Free tier includes 200 daily requests on free models; pay-as-you-go adds a 5% markup over base model costs with no subscription, seat fee, or minimum spend. | Pay-as-you-go is enabled in Requesty; Enterprise is priced through sales. | 2026-06-25 | Open |
| Glama | Global router | OpenAI / Anthropic / Google | Connects to 100+ models through an OpenAI-compatible API; the official page says there are no hidden fees and users pay listed prices. | Usage is tied to a Glama API key/account; billing details are handled in the account. | 2026-06-25 | Open |
| 接口AI | China relay | OpenAI / Claude / Gemini / DeepSeek / Qwen | The pricing page lists input, cache write, cache read, and output rates for Anthropic, OpenAI, and other models. | Pay-as-you-go with balance described as non-expiring; top-up and activation are handled in the dashboard. | 2026-06-25 | Open |
| 硅基流动 SiliconFlow | China relay | DeepSeek / Qwen / Kimi / GLM / Multimodal models | The pricing page is synchronized in real time and lists input, output, and cache-hit costs by provider and model. | Supports free start and pay-as-you-go; top-up and enterprise purchasing are handled through account or sales flows. | 2026-06-25 | Open |
| CloseAI | China relay | OpenAI / Claude / Gemini | Plans bill by forwarding multipliers over official API pricing: Business 1.5x, R&D 1.25x, Enterprise 1.1x; rare models may be higher. | R&D and Enterprise tiers have cumulative top-up thresholds; invoicing and corporate purchasing are supported. | 2026-06-25 | Open |
| 灵芽API | China relay | DeepSeek / Qwen / Kimi / GLM / Doubao | The model list shows support for 100+ models including DeepSeek, Qwen, Kimi, GLM, Doubao, and MiniMax. | Top-up and model rates should be checked in the Lingya API dashboard or model list. | 2026-06-25 | Open |
| LiteLLM | Self-hosted | OpenAI / Anthropic / Gemini / Bedrock / Azure | LiteLLM is a self-hosted or managed gateway; model pricing can be reviewed in its model catalog or configured through custom pricing. | Self-hosted deployments use the team's own upstream accounts; managed enterprise service requires contacting LiteLLM. | 2026-06-25 | Open |
| Portkey | Global router | OpenAI / Anthropic / Google / Mistral / Cohere | The model directory provides pricing data for 3000+ models across 39+ providers; Portkey Gateway offers a free tier and usage-based pricing. | Usage-based and Enterprise plans are enabled through Portkey account or sales. | 2026-06-25 | Open |
| Helicone | Global router | OpenAI / Anthropic / Gemini / Bedrock | Helicone provides AI Gateway, observability, and an API cost calculator; its gateway messaging emphasizes zero-markup pricing. | Billing depends on the Helicone account and connected upstream providers; exact plans are handled through Helicone account or sales. | 2026-06-25 | Open |