
Claude
From X ₽ / 1M tokensStreamingToolsReasoning
OpenAI, Anthropic and Kimi
through one API

From X ₽ / 1M tokensStreamingToolsReasoning

From X ₽ / 1M tokensStreamingToolsReasoning

From X ₽ / 1M tokensStreamingToolsReasoning
| Model | Input / 1M | Output / 1M |
|---|---|---|
| GPT-5.6 Sol | 31 ₽ | 186 ₽ |
| GPT-6 Astra | 62 ₽ | 310 ₽ |
| Claude Fable 5 | 298 ₽ | 1,488 ₽ |
| Kimi K3 | — | — |

Connect Relé to Codex and Claude Code.
Use a single API key across your coding workflow
and access supported AI models in one place.
No separate provider setup or extra accounts.
Just connect your key and start coding.
model =
model_provider = "rele"
[model_providers.rele]
base_url = "https://api.rele.ai/v1"
env_key = "RELE_API_KEY"
wire_api = "responses"Add Relé to your Codex or Claude Code configuration.
Set the endpoint and a single API key —
then switch between models
without changing your workflow.
One setup for all your requests.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.

OneAPI keeps access simple: one key, one balance and a predictable OpenAI-compatible interface.