> ## Documentation Index
> Fetch the complete documentation index at: https://docs.radium.cloud/llms.txt
> Use this file to discover all available pages before exploring further.

# Compatibility and limits

> What the Radium API supports, what it does not, and the limits that apply

# Compatibility and limits

Use this page when you need a quick answer about whether a specific OpenAI or Anthropic feature works on Radium.

## Supported and unsupported endpoints

| Endpoint | Status | Notes |
| - | - | - |
| `POST /v1/chat/completions` | Supported | Full OpenAI Chat Completions compatibility |
| `POST /v1/messages` | Supported | Full Anthropic Messages compatibility |
| `POST /v1/messages/count_tokens` | Supported | Anthropic-compatible token counting |
| `GET /v1/models` | Supported | Lists `hal-1.0`, `clarke-1.0`, `tycho-1.0` |
| `POST /v1/responses` | Not supported | Coming soon |
| `POST /v1/embeddings` | Not supported | Use a dedicated embedding provider |
| Batch | Not supported | — |
| File upload | Not supported | — |

## Parameter limits

| Parameter | Limit |
| - | - |
| `temperature` | 0 – 2 |
| `top_p` | 0 – 1 |
| `presence_penalty` | -2.0 – 2.0 |
| `frequency_penalty` | -2.0 – 2.0 |
| `max_tokens` | Model-dependent; see [Models](https://docs.radium.cloud/models/overview.md) |
| `seed` | Any integer |

## Tool calling

Tool calling works on `hal-1.0`, `clarke-1.0`, and `tycho-1.0` through both the OpenAI and Anthropic endpoints. The schema format, `tool_choice` behavior, and parallel tool execution match the OpenAI specification. There is no extra latency for tool-enabled requests.

## Structured outputs

JSON mode (`response_format: {type: "json_object"}`) and JSON Schema (`response_format: {type: "json_schema", ...}`) are supported on `hal-1.0` and `clarke-1.0`. Tycho 1.0 supports JSON mode. Structured outputs follow the same rules as OpenAI: the model may refuse if the prompt cannot satisfy the schema, and a `refusal` field may be present.

## Prompt caching

The `cache_control` parameter is supported on all models. Cache hits are billed at the cached input rate. There is no additional latency for writing or reading from cache.

## Vision limits

Vision input is supported on `hal-1.0` and `clarke-1.0`. Images can be sent as base64-encoded data URLs or public HTTP(S) URLs. See the [Multimodal images example](https://docs.radium.cloud/examples/multimodal-images.md) for code samples.

| Limit | Value |
| - | - |
| Max images per request | TBD — confirm with Engineering |
| Max image size | TBD — confirm with Engineering |
| Supported formats | JPEG, PNG, GIF, WebP |

Tycho 1.0 does not support vision input.

## Rate limits

Rate limits are enforced at the organization level and shared across all API keys.

| Limit | Value |
| - | - |
| Requests per minute (RPM) | 1,000 (org-wide) |
| Burst | 33 requests |
| Acceleration limit | Ratio 3.0 over 10-second windows |
| Credit reserved in flight | `retry-after` 3 s |
| Monthly spend limit | Set by you in Billing |
| Tokens per minute (TPM) | None in this phase |

See [Rate limits](https://docs.radium.cloud/core-concepts/rate-limits.md) for full details on error codes and backoff strategies.

## SDK compatibility

Any SDK or framework that can target the OpenAI Chat Completions API or the Anthropic Messages API works with Radium after changing the base URL and API key. Official SDKs, Claude Code, LangChain, LangGraph, Vercel AI SDK, and others have been verified.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.