# A2Agent

> A2Agent is an API gateway that exposes Chinese large language models (GLM, Kimi, DeepSeek, Qwen, MiniMax) through OpenAI-, Anthropic- and Gemini-compatible endpoints.

Last updated: 2026-09-20T07:49:44Z

## Quickstart

Base URL: https://api.a2agent.me
Authentication: send `Authorization: Bearer <your API key>` on every request.
Keys are created in the dashboard after signing up.

Compatible request surfaces:

- OpenAI chat completions: POST https://api.a2agent.me/v1/chat/completions
- OpenAI responses: POST https://api.a2agent.me/v1/responses
- OpenAI models: GET https://api.a2agent.me/v1/models
- Anthropic messages: POST https://api.a2agent.me/v1/messages
- Anthropic token counting: POST https://api.a2agent.me/v1/messages/count_tokens
- Gemini generate content: POST https://api.a2agent.me/v1beta/models/{model}:generateContent

Any client that speaks one of these three protocols works unchanged: point its base URL at this gateway and use an A2Agent key. That includes Claude Code (Anthropic surface), Codex CLI and Gemini CLI.

## Model pricing

All prices are USD per 1,000,000 tokens and already include any account-level rate multiplier.

| Platform | Model | Input (USD / 1M tokens) | Output (USD / 1M tokens) |
| --- | --- | --- | --- |
| deepseek | deepseek-v4-flash | 0.084 | 0.168 |
| deepseek | deepseek-v4-pro | 0.261 | 0.522 |
| glm | glm-5 | 0.6 | 1.92 |
| glm | glm-5.1 | 0.92 | 3.2 |
| glm | glm-5.2 | 0.84 | 2.64 |
| glm | glm-5.3 | 1.26 | 3.96 |
| glm | glm-5.3-flash | 0.135 | 0.45 |
| kimi | kimi-k2.5 | 0.42 | 2.1 |
| kimi | kimi-k2.7-code | 0.665 | 2.8 |
| kimi | kimi-k3 | 2.4 | 12 |
| minimax | MiniMax-M2.7 | 0.15 | 0.6 |
| minimax | minimax-m2.5 | 0.15 | 0.6 |
| minimax | minimax-m3 | 0.15 | 0.6 |
| qwen | qwen3.5-plus | 0.35 | 2.1 |
| qwen | qwen3.6-flash | 0.16 | 0.96 |
| qwen | qwen3.7-flash | 0.14 | 0.56 |
| qwen | qwen3.7-max | 1.36 | 4.08 |
| qwen | qwen3.7-plus | 0.96 | 3.84 |
| qwen | qwen3.8-flash | 0.135 | 0.423 |
| qwen | qwen3.8-max | 1.8 | 5.4 |

## Articles

Append `.md` to any article URL for its markdown source.

- [Closed vs. Open-Source LLMs: Current State, Capability Gap, and How to Choose](https://a2agent.me/blog/closed-vs-open-source-llms-2026) — 2026-09-18: Using Arena and Vals Index as a common yardstick, this post measures how far open-source LLMs trail closed models on user preference and real-task completion, and how cost should factor into choosing between them.
- [Picking an Agent Client After Ten Years in Engineering: Codex, OpenClaw, and AgentiLoop](https://a2agent.me/blog/mac-agent-harness-agentiloop-en) — 2026-09-14: Why Codex hides too much and OpenClaw stalls on deep tool chains, and what AgentiLoop/Agent does differently — AgentScript, privilege separation, Accessibility-based automation, and the harness details that decide whether a long task finishes.
- [From Handwritten Code to the Malleable Computer: DHH's Vision for the Agentic Era](https://a2agent.me/blog/from-handwritten-code-to-malleable-computer-dhh-agentic-era) — 2026-09-05: After reading Lex Fridman's long conversation with DHH, we ask a simple question: as AI gets better at building software, what will become truly scarce?
- [Your LLM Benchmark Is Probably Measuring the Wrong Reality](https://a2agent.me/blog/your-llm-benchmark-is-measuring-the-wrong-reality) — 2026-08-21: The AI Observatory puts 24,521 real conversations under one taxonomy and finds that AI usage changes with the dataset, the model, and the month. For anyone shipping an LLM product, that turns "which model is best?" into a much more interesting systems problem.
- [Kimi-K3's Frontend Win Complicates the U.S.-China AI Narrative](https://a2agent.me/blog/kimi-k3-frontend-arena-commentary) — 2026-07-21: Kimi-K3 tops the Frontend Code Arena at 1,679, briefly flipping the U.S.-China gap. Here is what the chart does — and does not — prove about the AI race.
- [One Key, Every Model: Welcome to A2Agent](https://a2agent.me/blog/welcome-to-a2agent) — 2026-06-15: A2Agent gives you a single OpenAI-compatible endpoint for top models across multiple providers — with transparent, pay-as-you-go pricing.
- [How to Integrate A2Agent with the OpenAI SDK](https://a2agent.me/blog/integrate-with-openai-sdk) — 2026-06-10: Already using the OpenAI SDK? Switching to A2Agent takes two lines — set the base URL and your A2Agent key.
- [New Models Added This Month](https://a2agent.me/blog/new-models-this-month) — 2026-06-05: We continually expand the catalog. Here are the latest additions and what they are good at.