Apmix
Apmix is an AI API aggregation platform that puts eleven model families — GPT, Claude, Gemini, Grok, DeepSeek, Qwen, GLM, Kimi, MiniMax, MiMo and Hunyuan — behind one API key and one monthly subscription from $4.99. Point Claude Code, Codex CLI, Cursor or your own code at api.apmix.ai using OpenAI- or Anthropic-compatible endpoints, and track a single weighted-token allowance instead of paying six providers separately.
What is Apmix?
Apmix is a low-cost AI API aggregator run by Nomvel that exposes the major model families through a single OpenAI- and Anthropic-compatible gateway at api.apmix.ai. Instead of holding separate accounts, keys and invoices with OpenAI, Anthropic, Google, xAI, DeepSeek and Alibaba, a developer creates one Apmix account, generates keys from the dashboard, and changes only the base URL inside tools already in use. The catalogue spans eleven families and 41 models: GPT, Claude, Gemini, Grok, DeepSeek, Qwen, GLM, Kimi, MiniMax, MiMo, Hunyuan and Cursor's Composer, with a models page and a model ranking page alongside the pricing table.
Access is sold as a subscription rather than prepaid pay-as-you-go tokens. Starter costs $4.99 a month for 60M weighted tokens, 13 models and 2 API keys; Pro is $14.99 for 200M weighted tokens, 28 models and 10 keys; Max is $29.99 for 500M weighted tokens, 38 models, unlimited keys and priority routing. Max covers the whole catalogue. Yearly billing keeps the same monthly allowance at a price that saves up to 12%, and a new account can start on a one-time grant of 4M free tokens shared across the free-tier models without a credit card.
Metering is where the product differs from an ordinary token bill. Every plan is a monthly allowance of weighted tokens shared across all models and all keys on the account. Each model carries a multiplier — 0.05× for MiMo V2.5, 0.5× for Grok and DeepSeek V4 Pro, 2× for Claude Sonnet 5, 8× for Claude Fable 5 — and prompt, completion and reasoning tokens are all multiplied by it before being deducted. The pricing page converts any plan's allowance into real tokens per model, and no daily or weekly caps apply unless the user sets them.
The service targets developers and small teams running coding agents or product features, so the homepage documents Claude Code and Codex CLI setup with base URL and key environment variables and notes that Cursor, official SDKs and curl work the same way. The dashboard shows the current month's weighted-token total, the projected percentage used and spend broken down by model. When an allowance runs out, requests return a clear error until the monthly reset, the next tier, or a one-time top-up of 10M to 100M weighted tokens from $1.49. Monthly plans are refundable in full below 5% usage, yearly plans are not refundable, and cancelling takes one click while keys keep working to the end of the paid period.
Apmix Core Features
Single key across families
one Apmix key reaches GPT, Claude, Gemini, Grok, DeepSeek, Qwen, GLM, Kimi, MiniMax, MiMo and Hunyuan.
OpenAI and Anthropic compatible
point your base URL at api.apmix.ai and keep the SDK, curl or CLI client you already use.
Weighted-token allowance
one monthly pool shared across every model and every key, with no daily or weekly caps by default.
Per-model multipliers
a published rate converts real prompt, completion and reasoning tokens into weighted tokens at 0.05× to 8×.
Tiered catalogue
13 models on Starter, 28 on Pro and 38 on Max, from cheap flash models up to frontier reasoning ones.
Usage top-ups
subscribers buy 10M, 30M, 60M or 100M extra weighted tokens from $1.49 without changing plan.
Dashboard usage meter
this month's weighted tokens, projected percentage used and spend broken down by model.
Who is Apmix for?
Coding-agent users are the primary audience. The homepage shows environment variables for Claude Code and Codex CLI before it explains anything else, which signals that Apmix is built for developers who run terminal agents all day and would rather pay one subscription from $4.99 to $29.99 than reconcile separate Anthropic, OpenAI and Google bills. Cursor users, and anyone comfortable editing a base URL in a config file, fall into the same group. Individual developers and side-project builders are the second fit. Starter's 60M weighted tokens across 13 models, plus 2 API keys and the 4M free tokens at sign-up, are sized for prototypes, personal automation and hobby apps where a fixed monthly price is easier to budget than metered credit. Small product teams belong here too. Pro adds Claude, Gemini and Qwen models with 10 API keys, so a team can share one allowance and one endpoint while every project keeps its own key. Teams that need frontier reasoning — Claude Opus, GPT-6 Astra, Kimi K3, Qwen3.8 Max — need Max, because those models sit in the highest tier. Agency and freelance developers serving several clients are another natural user: one subscription, one usage meter and one invoice are simpler to pass through than six provider accounts. People who want a consumer chat app are not the audience, because Apmix sells API access only. Nor is it aimed at teams that require a signed SLA, regional processing guarantees or self-hosted routing, since nothing on the site documents those.
Apmix Use Cases
Point Claude Code at api.apmix.ai so a single subscription covers your terminal coding sessions.
Run Codex CLI against the same key by switching the OpenAI base URL to apmix.
Swap a script from GPT to Claude by editing the model name and nothing else.
Replace six provider accounts and invoices with one monthly Apmix bill.
Prototype an app on the 4M sign-up tokens before paying for anything.
Buy a small top-up when a heavy month runs past the weighted-token allowance.
Check the dashboard's spend-by-model view to see which calls eat your allowance.
Compare per-model multipliers before pointing production traffic at a frontier model.
Apmix Pros and Cons
Pros
- Plans start at $4.99 a month for 60M weighted tokens, and yearly billing keeps the same allowance at up to 12% less.
- 4M free tokens at sign-up need no card, so the endpoint and clients can be tested before subscribing.
- OpenAI- and Anthropic-compatible routes mean Claude Code, Codex CLI and Cursor work by changing a base URL.
- One key, one usage meter and one invoice replace separate accounts and bills with each model provider.
- Monthly plans are refundable in full while under 5% of the allowance is used, and cancellation is a one-click action.
Cons
- Model access is tiered, so the $4.99 Starter plan excludes Claude, Gemini and Qwen models that only appear from Pro.
- Costs are expressed as multipliers rather than dollars, so estimating real spend means dividing your allowance by each model's rate.
- Frontier and reasoning models draw weighted tokens several times faster, and yearly plans are non-refundable once paid.
FAQ About Apmix
Apmix Pricing
Freemium: 4M free tokens at sign-up, then three monthly plans — Starter $4.99 for 60M weighted tokens, Pro $14.99 for 200M and Max $29.99 for 500M — with yearly billing saving up to 12% and one-time usage top-ups from $1.49.
Check official pricingFree
Every new account starts with 4M free tokens shared across the free-tier models (DeepSeek V4 Flash Free, DeepSeek V4.1 Flash Free and GPT 6 Luna Free), with no daily limits, no expiry and no card required.
Starter
60M weighted tokens a month, 13 models including Grok 4.x, DeepSeek V4, GLM-5.3 Flash, MiMo, MiniMax M3, Cursor Composer and GPT-5.6 Luna, plus 2 API keys and no daily or weekly caps.
Pro
200M weighted tokens a month, 28 models: everything in Starter plus Claude Haiku 4.5, Sonnet 5 and Sonnet 4.6, Gemini 3.x, Qwen3.8 Flash, Hunyuan 4 Preview, GLM-5.3, Kimi K2.7 Code and GPT-6 Sol, with 10 API keys.
Max
500M weighted tokens a month, 38 models: everything in Pro plus Claude Opus 5, GPT-5.6 Sol, GPT-5.5, Qwen3.8 Max, GPT-6 Astra, Claude Fable 5 and Kimi K3, with unlimited API keys and priority routing.
Apmix Alternatives
Ecommerce Data API
Amazon product, keyword, traffic and market research data through a REST API, remote MCP server and Agent Skill for developers and AI workflows.
FastRouter
FastRouter is a unified LLM gateway that gives developers one OpenAI-compatible API for 200+ models from providers such as OpenAI, Anthropic, Google, and xAI. It adds automatic model routing, fallbacks, caching, spend caps, role-based access, guardrails, and per-request observability, while passing provider token prices through with no markup on any plan, including the free tier.
MapGO
How far is this point from the sea? From the nearest border? Which country, region and district are it in? MapGO answers those in one click.