The Cheapest AI Model API
DeepSeek, GLM, Kimi, Qwen & MiniMax below official prices. One API key. Pay as you go.
All Major Models
GLM, Kimi, DeepSeek, Qwen and MiniMax with one API key. Switch models by changing the model name.
OpenAI & Anthropic Compatible
Works with the OpenAI and Anthropic SDKs. Change the base URL and API key, keep your code.
Security & Privacy
Every request goes over HTTPS. Our privacy policy covers how we process and retain data.
Cheapest AI models right now
The 10 cheapest AI models on A2Agent by blended price (input and output weighted 3:1), biggest discount first. Our price is what your API key pays; the official price is the vendor list price.
| Model | Our price (in / out) | Official price (in / out) | You save |
|---|---|---|---|
| MiniMax M3 | $0.150 / $0.600 | $0.300 / $1.20 | -50% |
| DeepSeek V4 Flash | $0.084 / $0.168 | $0.140 / $0.280 | -40% |
| DeepSeek V4.1 Flash | $0.162 / $0.486 | $0.270 / $0.810 | -40% |
| DeepSeek V4 Pro | $0.261 / $0.522 | $0.435 / $0.870 | -40% |
| GLM-5.3 Flash | $0.105 / $0.350 | $0.150 / $0.500 | -30% |
| Qwen3.7 Flash | $0.140 / $0.560 | $0.200 / $0.800 | -30% |
| Qwen3.6 Flash | $0.140 / $0.840 | $0.200 / $1.20 | -30% |
| Qwen3.5 Plus | $0.350 / $2.10 | $0.500 / $3.00 | -30% |
| Qwen3.8 Flash | $0.135 / $0.423 | $0.150 / $0.470 | -10% |
| MiniMax M3.1 Flash | $0.135 / $0.540 | $0.150 / $0.600 | -10% |
Works with leading LLMs and AI coding clients
GLM
Kimi
DeepSeek
Qwen
MiniMax
Claude Code
Codex
OpenCode
Aider
Gemini CLI
CodeBuddy
GitHub Copilot
Factory Droid
TRAE
Goose
Crush
Cline
Roo Code
Kilo Code
Continue
JetBrains AI
Cursor
Zed
QoderGet started in four steps
From sign-up to your first API call — no complex setup.
Sign up
Create an account with your email address.
Choose a model
Choose a model from DeepSeek, GLM, Kimi, Qwen, or MiniMax.
Get an API key
Sign in to the dashboard and create your API key.
Start Building
Add credits, follow your client setup guide and verify a small API request.
Pay only for what you use
Developer
Pay as you go
- Pay as you go with no monthly fee
- Full GLM / Kimi support
- DeepSeek deep-thinking support
- Qwen / MiniMax support
- Balance never expires
- For individual developers and students
Enterprise
Volume discount
- Volume top-up discounts
- Dedicated concurrency limits
- Higher request-per-minute limits
- Every model and provider channel on the platform
- Balance never expires
- For SaaS companies & teams
Supported Payment Methods
Top-ups and service activation are available. Final payment methods are shown at checkout.
Why Choose A2Agent?
Connect your app, track usage, and manage costs.
Competitive Pricing
Compare input and output rates in the model catalog. Discounts depend on the model and API key group.
Check the rate for your workloadPay as you go
Pay-as-you-go has no recurring commitment. Top up from $1 with no top-up fee.
Only pay for what you useCompatible APIs
OpenAI, Anthropic and Gemini endpoints are available. Client features and model support vary.
Migrate existing projects with minimal code changesJoin the A2Agent AI developer community
Share what you are building and get help connecting models to your workflow.
Real Workflow Breakdowns
See examples of script automation, MCP integrations, RAG, and agent workflows.
Help with integration issues
Ask the community about setup errors, rate limits, context limits, and tool calls.
Discuss models and tools
Discuss long context, image input, reasoning, coding agents, and evaluations.
Share projects and feedback
Compare notes with developers building AI products, automation, and coding assistants.
Join us on Telegram or Discord for model updates, project examples, and technical discussions.
Frequently Asked Questions
What is the cheapest AI model API?
The ten cheapest models on A2Agent are in the table at the top of this page, picked by blended price (input and output weighted 3:1) and listed biggest discount first. Flash tiers such as DeepSeek V4 Flash, GLM-5.3 Flash and the Qwen Flash models are usually among them. The prices are live. Billing is per token and no subscription is required.
What is the cheapest LLM for coding?
Open the Models page and filter by Coding. The table lists the cheapest model first, so the top row is the lowest-priced model with the coding tag. A wrong answer is billed the same as a right one, so run the same task on two or three candidates in Claude Code or Codex CLI before you pick a default.
Is there a free AI API key?
Yes. Every new account starts with $0 in free credit, no card required, and it works on any model your key can access. When it runs out you pay as you go, from $1 a top-up. Free credit and top-up bonus credit cannot be refunded.
Does A2Agent work with Claude Code and Codex?
Yes. Claude Code talks to the Anthropic Messages endpoint (/v1/messages) and Codex CLI to OpenAI Responses (/v1/responses). Both take an A2Agent key and a model ID from the Models page. The Integrations page has a setup guide for each, along with Cline and 30 more tools.
How does A2Agent compare to OpenRouter?
OpenRouter is the wider catalog: 500+ models from many providers, with provider prices passed through. It charges a 5.5% fee ($0.80 minimum) when you buy credits by card, and unused credits may expire after 365 days (OpenRouter FAQ and terms, checked September 28, 2026). A2Agent carries five families, DeepSeek, GLM, Kimi, Qwen and MiniMax, priced below the vendors’ official list prices, with no top-up fee and credits that do not expire while your account is active. The Pricing page compares every shared model side by side.
Are these the official models?
Yes. Every model comes straight from its vendor’s official platform, and A2Agent does not train or modify any of them. Each model ID names the release it serves, such as deepseek-v4-flash or glm-5.3. Failover happens between A2Agent’s own accounts with that vendor, never to a third-party host, so an ID always means the same model.
Does A2Agent have GPT, Claude or Gemini models?
No. The catalog is DeepSeek, GLM, Kimi, Qwen and MiniMax. A2Agent speaks the OpenAI, Anthropic and Gemini protocols, so clients built for those APIs connect, but the model behind the call is always one of those five families. Keep a second provider if you also need GPT, Claude or Gemini.
Do I need accounts with DeepSeek, Zhipu, Moonshot, Alibaba or MiniMax?
No. You sign up at A2Agent with an email address and pay from one balance for all five families. There is no vendor registration and no Chinese phone number to supply. A2Agent holds the vendor accounts and bills you per token.
Try a model in your app
Check the rate, create a key and verify your first request.
