#参数错误文案
本页按源码生成，列出输入校验的显式英文文案和参数模板；不包含供应商原始错误或内部异常。 通常返回 400 / invalid_request_error 或 unsupported_parameter；部分 Messages 原生控制拒绝没有独立 code。 同一文案仅在相应模型能力、参数组合或协议转换路径中触发，不表示所有模型都有这一限制。 param 为变量或表达式时显示源码模板；未声明时由外层入口设置字段。第三方校验器的动态字段文案不能穷尽。 原生支持的协议优先使用原生路径，转换校验不等于所有请求都默认转换。 错误码与处理建议见 网关错误。
| 英文 message 原文或参数模板 | param / 字段模板 |
|---|---|
| Anthropic cache_control blocks are not supported by TokenPostie Messages V1 yet. | {value}.{index}.cache_control |
| Anthropic content block at {value}.{index} must be an object. | {value}.{index} |
| Anthropic content block type '{type}' is not supported by TokenPostie Messages V1 yet. | {value}.{index}.type |
| Anthropic message at index {index} has unsupported role '{role}'. | messages.{index}.role |
| Anthropic message at index {index} must be an object. | messages.{index} |
| Anthropic Messages request body must be a JSON object. | body |
| Anthropic Messages request requires a non-empty model. | model |
| Anthropic Messages request requires a positive integer max_tokens. | max_tokens |
| Anthropic Messages request requires at least one message. | messages |
| Anthropic Messages V1 only supports string content or an array of text blocks. | {field} |
| Anthropic parameter '{value}' is not supported by TokenPostie Messages V1 yet. | {field} |
| Anthropic text content block array must contain at least one text block. | {field} |
| Anthropic text content blocks require a string text field. | {value}.{index}.text |
| background responses are not supported by this gateway. | background |
| conversation is not supported by this stateless gateway. | conversation |
| Converted reasoning input requires plaintext summary_text parts. | {field}.summary |
| Embedding request body is invalid. | body |
| Encrypted reasoning state cannot be converted to Chat. Use a native Responses route. | {field}.encrypted_content |
| Null content requires assistant tool calls | 由入口确定 |
| Only assistant messages may contain tool calls | 由入口确定 |
| Output token limit exceeds the configured model maximum of {max}. | {field} |
| previous_response_id is not supported by this stateless gateway. Resend prior output items in input. | previous_response_id |
| Provide input or instructions | 由入口确定 |
| Rerank request body is invalid. | body |
| Responses API background must be a boolean. | background |
| Responses API content part must be an object. | {field} |
| Responses API function tool requires name. | {field}.name |
| Responses API function tool_choice requires name. | tool_choice.name |
| Responses API function_call arguments must be a JSON string. | {field}.arguments |
| Responses API function_call requires call_id. | {field}.call_id |
| Responses API function_call requires name. | {field}.name |
| Responses API function_call_output requires call_id. | {field}.call_id |
| Responses API input array must contain at least one item. | input |
| Responses API input item type must be a string. | {field}.type |
| Responses API input must be a non-empty string or an array of input items. | input |
| Responses API input string must not be empty. | input |
| Responses API input_image currently requires image_url. file_id images are not supported. | {field}.image_url |
| Responses API instructions must be a string. | instructions |
| Responses API json_schema format requires name and schema. | text.format.name / text.format.schema |
| Responses API max_output_tokens must be a positive integer. | max_output_tokens |
| Responses API message content must be a string or an array of content parts. | {field} |
| Responses API message role must be developer, system, user, or assistant. | {field}.role |
| Responses API parallel_tool_calls must be a boolean. | parallel_tool_calls |
| Responses API reasoning must be an object. | reasoning |
| Responses API reasoning.effort must be a non-empty string. | reasoning.effort |
| Responses API request body must be a JSON object. | body |
| Responses API request requires a non-empty model. | model |
| Responses API store must be a boolean. | store |
| Responses API stream must be a boolean. | stream |
| Responses API temperature must be a finite number. | temperature |
| Responses API text content part requires text. | {field}.text |
| Responses API text must be an object. | text |
| Responses API text.format requires a type. | text.format.type |
| Responses API tool must be an object. | {field} |
| Responses API tool_choice must be auto, none, required, or a named function. | tool_choice |
| Responses API tools must be an array. | tools |
| Responses API top_p must be a finite number. | top_p |
| Responses API user must be a string. | user |
| Responses requests must use max_output_tokens, not Chat token limit parameters. | max_output_tokens |
| Specify only one output token limit | 由入口确定 |
| stored responses are not supported by this stateless gateway. Use store=false. | store |
| Temperature must be between 0 and 2. | temperature |
| The selected model does not apply top_k. | top_k |
| The selected model does not enforce thinking.budget_tokens. Use output_config.effort and max_tokens instead. | thinking.budget_tokens |
| The selected model does not generate reasoning summaries. | reasoning.summary |
| The selected model does not support forced tool_choice while thinking is enabled. Explicitly disable thinking, or use tool_choice auto. | tool_choice |
| The selected model does not support the requested reasoning effort. | {field} |
| The selected model only supports output_config.effort. | output_config.{value} |
| The selected model supports enabled or disabled thinking, not adaptive thinking. | thinking.type |
| Thinking mode and reasoning effort must not conflict. | {field} |
| This model requires a native protocol route; protocol conversion is disabled. | model |
| This route does not support native Messages effort control. | output_config |
| Tool messages require tool_call_id | 由入口确定 |
| top_p must be greater than 0 and at most 1. | top_p |