DocumentationModelsBack to top
Skip to content
Models
Rebyte Docs
Agents API
OverviewQuickstartRebyte Agent SDKUpgrade to 0.3.0OpenAI compatibilityArchitectureModelsConfiguring Agents
Sessions
Run and continue sessionsEvents and conversation historyManage sessionsWebhooks
Environments and sandboxes
Rebyte SandboxesSandbox lifecycleSandbox securityFiles and artifacts
Tools and integrations
Dynamic WorkflowWeb searchFunctionsMCP connectionsPlatform connectionsPluginsVaults
Workflow AgentsScheduled AgentsMulti-agentObservability and usageTracing

Not yet supported

Developer documentation/Agents API

Models

Supported Agent model IDs, their upstream routes, and migration from retired names.

Agents API2 min read
On this pageMoving from retired model IDs

Use the exact model ID below when creating or updating an Agent. The same catalog is used by the web app, iOS, and Agent API. The API requires an explicit model when defining an Agent; new product UI Agents start with deepseek-flash.

Model IDModelProvider route
deepseek-flashDeepSeek V4.1 FlashDeepSeek · deepseek-flash
claude-sonnet-5Claude Sonnet 5Anthropic · claude-sonnet-5
claude-opus-5Claude Opus 5Anthropic · claude-opus-5
gemini-3.8-flashGemini 3.8 FlashOpenRouter · google/gemini-3.8-flash
glm-5.3GLM 5.3OpenRouter · z-ai/glm-5.3
qwen3.8-maxQwen 3.8 MaxOpenRouter · qwen/qwen3.8-max
kimi-k3Kimi K3OpenRouter · moonshotai/kimi-k3

The existing GPT choices are gpt-6-sol and gpt-6-luna. Their routing and access policy are unchanged by this catalog update.

Moving from retired model IDs

From October 2, 2026, new Agent API requests must use the current non-GPT IDs. Retired IDs are rejected as unsupported models; they do not silently route to a different model. Update model strings in your application and agent.toml. Existing saved Agents and their executable Session configurations have been migrated as follows:

Previous model familyCurrent model ID
deepseek-v4-flash, deepseek-v4-prodeepseek-flash
Claude Sonnet 4.xclaude-sonnet-5
Claude Opus 4.xclaude-opus-5
gemini-3.5-flashgemini-3.8-flash
GLM 5, 5.1, 5.2, 5 Turboglm-5.3
Qwen3 Max variantsqwen3.8-max

These changes apply to the Agent model. Claude Code and Codex model settings are separate and remain unchanged. Historical usage and charges retain their original model records.

deepseek-flash uses DeepSeek's explicit Flash route. Rebyte's published Flash rates are $0.45 per million uncached input tokens, $0.009 per million cached input tokens, and $1.80 per million output tokens. These are fixed Rebyte rates, based on DeepSeek's peak list prices with Rebyte's proxy margin; they do not vary with upstream off-peak discounts.

Model access depends on organization credits and subscription policy.

Send OpenAI-Beta: agents=v1 with raw HTTP requests. The official OpenAI Agents SDK sets this header automatically. See Configuring Agents for saved Agents, inline definitions, and Session model updates.

Protocol referenceOpenAI Agents API ↗