Edit Model failed to apply, but just paste its reasoning process to my code.

[Z2IRIM]

Before submitting your bug report

• I've tried using the "Ask AI" feature on the Continue docs site to see if the docs have an answer
• I'm not able to find a related conversation on GitHub discussions that reports the same bug
• I'm not able to find an open issue that reports the same bug
• I've seen the troubleshooting guide on the Continue Docs

Relevant environment info

- OS: Win11.  Vscode Extension
- Continue version:I tried both Continue 1.12.7 and prelease version.
- IDE version: VSCode 1.11.0
- Model: gpt-5.4-mini gpt-5.3-codex
- config:
  
name: Customize1
description: Customize1
version: 1.0.0
schema: v1
models:
  - name: gpt-5.4-mini
    provider: openai
    model: gpt-5.4-mini
    apiKey: x
    apiBase: x
    roles:
      - chat
      - edit
    requestOptions:
      extraBodyProperties:
        think: false
    chatOptions:
      baseAgentSystemMessage: |
        You are a coding agent.
        Use tools when needed.
        Prefer minimal tool calls.
        Multiple read-only tools can be used in parallel.
        Answer directly.In Chinese.
        Constraints:
        - Disable Thinking
        - Max 5 bullets
        - No filler
        - No repetition

  - name: gpt-5.3-codex
    provider: openai
    model: gpt-5.3-codex
    apiKey: x
    apiBase: x
    roles:
      - edit

  - name: qwen3-coder-plus
    provider: openai
    model: qwen3-coder-plus
    apiBase: x
    apiKey: x
    roles:
      - chat

  - name: Codestral
    provider: mistral
    model: codestral-latest
    apiBase: x
    apiKey: x
    roles:
      - autocomplete
    autocompleteOptions:
      maxPromptTokens: 2048
      onlyMyCode: true

  - name: MorphFastApply
    provider: openai
    model: morph/morph-v3-fast
    apiKey: x
    apiBase: x
    roles:
      - apply
    promptTemplates:
      apply: |
        Original: {{{original_code}}}
        New: {{{new_code}}}

        Please generate the final code without any markers or explanations.
  
  OR link to agent in Continue hub:

Description

Edit Model just paste its first-two line in his response into my code, and repeat twice. I tried both gpt5.4 and 5.3 codex model and get the same result. I tried removed rules and prompt file, and still.

To reproduce

No response

Log output

Prompt Log:
system
## Output
- Be concise
- No filler
- No repetition
- Direct answer only

## Code blocks
- Always include language + file path
- Format: <lang> <path>
- No markdown outside code unless needed

## Code generation
- Code only
- No explanation
- No comments unless required

## Large code
- Use placeholders for unchanged parts
- Keep only modified sections explicit

## Editing
- Input Code Only if you are in Edit Mode
- Do not output full files unnecessarily
- Prefer minimal diffs

## Reasoning
- Do not use thinking
- No thinking text
- No chain-of-thought

## Behavior
- Do not output placeholders like "uncertain".
- Always produce a best-effort code result.
user
The user has requested a section of code in a file to be rewritten.

This is the prefix of the file:
:Untitled-1
import { IoIosArrowDropdown } from "react-icons/io";
const SkuSelector = ({ placeholder, options }) => {
  return (



This is the suffix of the file:
:Untitled-1

        {/*top-1/2 + transform -translate-y-1/2 可以控制控件位置纵向居中*/}
        <IoIosArrowDropdown className="h-4 w-4 text-apple-text-light dark:text-apple-text-dark" />
      </div>
    </div>
  );
};

export default SkuSelector;



This is the code to rewrite:
:Untitled-1
    <div className="relative w-full min-w-0">
      <select className="w-full min-w-0 appearance-none rounded-md border border-apple-gray-200 bg-none bg-no-repeat px-3 py-2 pr-8 text-sm text-apple-text-light dark:border-apple-gray-800 dark:text-apple-text-dark truncate overflow-hidden whitespace-nowrap">
        <option value="">{placeholder}</option>
        {options.map((option) => (
          <option key={option} value={option}>
            {option}
          </option>
        ))}
      </select>
      <div className="absolute right-2 top-1/2 -translate-y-1/2">


The user's request is: "缩小字号"

DO NOT output any natural language, only output the code changes.

Here is the rewritten code:
 Options
thinking
**Adjusting code for font size**

I see that the request "缩小字号" translates to reducing the font size. I’ll likely just need to make minimal
thinking
 changes in the code. Specifically, I'll switch classes from text-sm to perhaps text-xs to achieve that smaller size. 

I'll prepare the entire modified snippet in a code
thinking
 block, including the language and file path as instructed but won’t include any explanations. Let’s focus on just crafting that snippet!**Adjusting code for font size**

I see that the request "缩小字号" translates to reducing the font size. I’ll likely just need to make minimal changes in the code. Specifically, I'll switch classes from text-sm to perhaps text-xs to achieve that smaller size. 

I'll prepare the entire modified snippet in a code block, including the language and file path as instructed but won’t include any explanations. Let’s focus on just crafting that snippet!
assistant
 Untitled
assistant
-1
    <div className="relative w-full min-w-0">
      <select className="w-full min-w-0 appearance-none rounded
assistant
-md border border-apple-gray-200 bg-none bg-no-repeat px-3 py-2 pr-8 text-xs text-apple-text-light dark:border-
assistant
apple-gray-800 dark:text-apple-text-dark truncate overflow-hidden whitespace-nowrap">
        <option value="">{placeholder}</option>
        {options.map((option
assistant
) => (
          <option key={option} value={option}>
            {option}
          </option>
        ))}
      </select>
      <div class
assistant
Name="absolute right-2 top-1/2 -translate-y-1/2">


Output Log:
thinking
**Adjusting code for font size**

I see that the request "缩小字号" translates to reducing the font size. I’ll likely just need to make minimal
thinking
 changes in the code. Specifically, I'll switch classes from text-sm to perhaps text-xs to achieve that smaller size. 

I'll prepare the entire modified snippet in a code
thinking
 block, including the language and file path as instructed but won’t include any explanations. Let’s focus on just crafting that snippet!**Adjusting code for font size**

I see that the request "缩小字号" translates to reducing the font size. I’ll likely just need to make minimal changes in the code. Specifically, I'll switch classes from text-sm to perhaps text-xs to achieve that smaller size. 

I'll prepare the entire modified snippet in a code block, including the language and file path as instructed but won’t include any explanations. Let’s focus on just crafting that snippet!
assistant
 Untitled
assistant
-1
    <div className="relative w-full min-w-0">
      <select className="w-full min-w-0 appearance-none rounded
assistant
-md border border-apple-gray-200 bg-none bg-no-repeat px-3 py-2 pr-8 text-xs text-apple-text-light dark:border-
assistant
apple-gray-800 dark:text-apple-text-dark truncate overflow-hidden whitespace-nowrap">
        <option value="">{placeholder}</option>
        {options.map((option
assistant
) => (
          <option key={option} value={option}>
            {option}
          </option>
        ))}
      </select>
      <div class
assistant
Name="absolute right-2 top-1/2 -translate-y-1/2">

[oyster-lab]

same problem

[Z2IRIM]

same problem
hey dude, I just fix it. Go edit your config.yaml and disable reasoning for your edit model. This is mine if you are using OpenAI compatible model you can copy it to the bottom of your model config block. otherwise you should check your provider docs.

requestOptions:
extraBodyProperties:
reasoning:
effort: none