Enable Continue Plugin Causes Significant Slowdown in Python Inline Suggestions

[Driftcell]

Before submitting your bug report

• I believe this is a bug. I'll try to join the Continue Discord for questions
• 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: Windows 11 Pro (Version 24H2, OS Build 26100.3194)
Continue version: 1.0.2
IDE version: Visual Studio Code 1.97.2
Model: deepseek v3 official

{
  "models": [
    {
      "title": "Deepseek",
      "model": "deepseek-chat",
      "provider": "deepseek",
      "apiKey": "sk-xxx"
    }
  ],
  "tabAutocompleteModel": {
    "title": "Deepseek",
    "model": "deepseek-chat",
    "provider": "deepseek",
    "apiKey": "sk-xxx"
  },
  "contextProviders": [
    {
      "name": "code",
      "params": {}
    },
    {
      "name": "docs",
      "params": {}
    },
    {
      "name": "diff",
      "params": {}
    },
    {
      "name": "terminal",
      "params": {}
    },
    {
      "name": "problems",
      "params": {}
    },
    {
      "name": "folder",
      "params": {}
    },
    {
      "name": "codebase",
      "params": {}
    }
  ],
  "slashCommands": [
    {
      "name": "share",
      "description": "Export the current chat session to markdown"
    },
    {
      "name": "cmd",
      "description": "Generate a shell command"
    },
    {
      "name": "commit",
      "description": "Generate a git commit message"
    }
  ],
  "embeddingsProvider": {
    "provider": "free-trial"
  },
  "reranker": {
    "name": "free-trial"
  }
}

Description

After enabling the Continue plugin in VS Code, I noticed a significant performance degradation in Python inline suggestions. Normally, inline suggestions (e.g., for basic constructs like async def pass) appear almost instantly. However, with the Continue plugin enabled, it takes approximately 5 seconds for these suggestions to appear.

This issue occurs consistently across different Python files and projects. Disabling the Continue plugin immediately resolves the slowdown, and inline suggestions return to their normal speed.

To reproduce

1. Install the Continue plugin in VS Code.
2. Open a Python file.
3. Start typing a basic Python construct, such as async def pass.
4. Observe the delay in inline suggestions.

Log output

==========================================================================
==========================================================================
##### Completion options #####
{
  "contextLength": 8096,
  "model": "deepseek-chat",
  "maxTokens": 4096,
  "temperature": 0.01,
  "stop": [
    "<｜fim▁begin｜>",
    "<｜fim▁hole｜>",
    "<｜fim▁end｜>",
    "//",
    "<｜end▁of▁sentence｜>",
    "/src/",
    "#- coding: utf-8",
    ""
  ]
}

##### Prompt #####
Prefix: 
# login/goofish.py
import asyncio

class GoofishLoginHelper:
    def __init__(self):
        ...

    async def login(self):
        ...

    a
Suffix: 
==========================================================================
==========================================================================
##### Completion options #####
{
  "contextLength": 8096,
  "model": "deepseek-chat",
  "maxTokens": 4096,
  "temperature": 0.01,
  "stop": [
    "<｜fim▁begin｜>",
    "<｜fim▁hole｜>",
    "<｜fim▁end｜>",
    "//",
    "<｜end▁of▁sentence｜>",
    "/src/",
    "#- coding: utf-8",
    ""
  ]
}

##### Prompt #####
Prefix: 
# login/goofish.py
import asyncio

class GoofishLoginHelper:
    def __init__(self):
        ...

    async def login(self):
        ...

    
Suffix: 
==========================================================================
==========================================================================
##### Completion options #####
{
  "contextLength": 8096,
  "model": "deepseek-chat",
  "maxTokens": 4096,
  "temperature": 0.01,
  "stop": [
    "<｜fim▁begin｜>",
    "<｜fim▁hole｜>",
    "<｜fim▁end｜>",
    "//",
    "<｜end▁of▁sentence｜>",
    "/src/",
    "#- coding: utf-8",
    ""
  ]
}

##### Prompt #####
Prefix: 
# login/goofish.py
import asyncio

class GoofishLoginHelper:
    def __init__(self):
        ...

    async def login(self):
        ...

    a
Suffix: 
==========================================================================
==========================================================================
##### Completion options #####
{
  "contextLength": 8096,
  "model": "deepseek-chat",
  "maxTokens": 4096,
  "temperature": 0.01,
  "stop": [
    "<｜fim▁begin｜>",
    "<｜fim▁hole｜>",
    "<｜fim▁end｜>",
    "//",
    "<｜end▁of▁sentence｜>",
    "/src/",
    "#- coding: utf-8",
    ""
  ]
}

##### Prompt #####
Prefix: 
# login/goofish.py
import asyncio

class GoofishLoginHelper:
    def __init__(self):
        ...

    async def login(self):
        ...

    
Suffix: 
==========================================================================
==========================================================================
Completion:
async def logout(self):
        ...

[Driftcell]

I have tested this issue repeatedly and confirmed that the slowdown is directly caused by the Continue plugin. When I uninstall the Continue plugin, the inline suggestion speed returns to normal immediately. Reinstalling the plugin brings back the slowdown. I am using the official Python extension pack for VS Code.

[Symfomany]

What's extensions VSCode do you use in your Python ? Version of Python ? Can you switch another language to try if is LLM ? Distill / R1 / V3 of DeepSeek ?

[Driftcell]

My environment is Python 3.13, using the official Python extension suite (including Pylance, isort, etc.), and calling the model via the official Deepseek API. Although the response time of the Deepseek API is slow, it typically does not significantly impact inline suggestions.

I tested the Continue plugin in a TypeScript project, and the inline suggestions performed normally without noticeable delays. However, the issue persists in Python projects. The situation is slightly better today compared to yesterday, but the delay is still uncomfortable, especially when typing common Python syntax structures (e.g., async def), where the response time for inline suggestions is noticeably longer.

[luxixing]

Change the completion model to a lower-latency API.
The issue may not be a bug but simply the model API responding too slowly.

[A-Yaskevich]

@Driftcell can you please check if slowdown of inline suggestions happen only when there is a plenty of uncommited code?