Use Mimxo AI with OpenAI Codex CLI
Codex CLI supports custom model providers. Mimxo AI implements the OpenAI Responses API, so Codex works with its default wire protocol.
Configure ~/.codex/config.toml
model = "gpt-6.1-sol" model_provider = "mimxo" [model_providers.mimxo] name = "Mimxo AI" base_url = "https://mimxo.com/v1" env_key = "MIMXO_API_KEY" wire_api = "responses"
export MIMXO_API_KEY=sk-your-mimxo-key codex
If a model does not support the Responses API, set wire_api = "chat" to use Chat Completions.
Recommended models
| Model | Provider | Input / 1M | Output / 1M | Cache read / 1M | List input / output |
|---|---|---|---|---|---|
qwen3.7-max | Alibaba Qwen | $0.4425 | $1.327 | $0.0885 | $1.475 / $4.425 |
claude-opus-5 | Anthropic | $1.5 | $7.5 | $0.15 | $5 / $25 |
claude-sonnet-5 | Anthropic | $0.6 | $3 | $0.06 | $2 / $10 |
deepseek-v4.1-flash | DeepSeek | $0.009 | $0.15 | $0.00769 | $0.03 / $0.5 |
gemini-3.1-pro-preview | $0.6 | $3.6 | $0.06 | $2 / $12 | |
kimi-k2.7-code | Moonshot Kimi | $0.2014 | $1.005 | $0.054 | $0.6712 / $3.35 |
gpt-6.1-sol | OpenAI | $0.6 | $3 | $0.03 | $2 / $10 |
glm-5.3 | Zhipu GLM | $0.42 | $1.32 | $0.042 | $1.4 / $4.4 |
Prices in USD per 1M tokens, read live from /api/pricing. "List" is the reference list price; Mimxo AI bills the first columns.
FAQ
Which models work best with Codex?
gpt-6.1-sol, gpt-6-sol and gpt-5.5 via the Responses API. Other models can be used with wire_api = "chat".
How much does it cost?
Mimxo AI bills every model at 0.3× list price per 1M tokens. Top-ups start at $1.
Other tools: OpenCode · Claude Code · Cursor · OpenClaw · Aider · Cline / Roo Code