Models

Vikasit Code ships with 5 curated AI models, all hosted in India by Chandorkar Technologies. Every model is available on every subscription plan.

---

Available models

Model	Model ID	Multiplier	Best for
Flash	vikasit/flash	1x	Quick tasks, lightweight completions
Standard/3	vikasit/std3	3x	General-purpose coding, balanced quality
Coder	vikasit/coder	7x	Code generation, refactoring, debugging
Max	vikasit/max	15x	Complex reasoning, architecture decisions
Fast	vikasit/fast	18x	High-speed inference for demanding tasks

The multiplier determines how quickly a model consumes your rate limit budget. A 1x model uses the base rate, while an 18x model uses 18 times the base rate per request. See rate limits for details.

---

Flash

The lightest model in the Vikasit lineup. Flash is designed for quick, low-cost tasks like generating boilerplate, simple completions, and rapid iteration.

• Multiplier: 1x
• Use when: You need fast responses for straightforward tasks
• Model ID: vikasit/flash

---

Standard/3

A well-rounded model that balances quality and speed. Standard/3 handles most everyday coding tasks effectively.

• Multiplier: 3x
• Use when: You need reliable quality for general development work
• Model ID: vikasit/std3

---

Coder

Purpose-built for code generation. Coder excels at writing new features, refactoring existing code, and debugging issues.

• Multiplier: 7x
• Use when: You are building features, fixing bugs, or refactoring code
• Model ID: vikasit/coder

---

Max

The most capable model for complex reasoning tasks. Max handles architecture decisions, intricate multi-file changes, and nuanced problem-solving.

• Multiplier: 15x
• Use when: You need deep reasoning for complex tasks
• Model ID: vikasit/max

---

Fast

High-speed inference for demanding workloads. Fast delivers quick results on tasks that would normally require significant compute.

• Multiplier: 18x
• Use when: You need rapid responses on computationally intensive tasks
• Model ID: vikasit/fast

---

Select a model

Once you have configured your account, you can select a model by running:

/models

This will show all 5 available models. Select the one that fits your task.

---

Set a default

To set a default model, add the model key to your Vikasit Code config.

vikasit.json

{
  "$schema": "https://vikasit.ai/config.json",
  "model": "vikasit/coder"
}

The format is vikasit/<model_id>.

---

Small model

Vikasit Code uses a separate lightweight model for background tasks like generating session titles. You can configure this with the small_model option.

vikasit.json

{
  "$schema": "https://vikasit.ai/config.json",
  "small_model": "vikasit/flash"
}

By default, Vikasit Code uses Flash for these tasks.

---

Loading models

When Vikasit Code starts up, it checks for models in the following priority order:

1. The --model or -m command line flag.

   vikasit --model vikasit/coder

2. The model key in the Vikasit Code config.

   vikasit.json

   {
     "$schema": "https://vikasit.ai/config.json",
     "model": "vikasit/coder"
   }

3. The last used model.

4. The first model using an internal priority.

---

Plan and access

All 5 models are available on every subscription plan (Lite, Pro, and Max). The difference between plans is the rate limit budget, not which models you can access. Learn more about plans.