feat: add ModelsLab as LLM provider

[adhikjoshi]

Summary

Adds ModelsLab as an OpenAI-compatible LLM provider, following the same pattern as the existing OpenRouter and Novita integrations.

ModelsLab provides access to 10,000+ open-source models (Llama 3, Mistral, DeepSeek, and more) through an OpenAI-compatible chat endpoint.

API docs: https://docs.modelslab.com
API key: https://modelslab.com/account/api-key
Endpoint: https://modelslab.com/api/uncensored-chat/v1

Changes

File	What
application/llm/modelslab.py	ModelsLabLLM class extending OpenAILLM
application/llm/llm_creator.py	Register "modelslab" provider
application/core/settings.py	Add MODELSLAB_API_KEY env var
application/core/model_settings.py	ModelProvider.MODELSLAB + _add_modelslab_models()
application/core/model_configs.py	MODELSLAB_MODELS (Llama 3 8B/70B, Mistral 7B)
application/core/model_utils.py	Add modelslab to provider_key_map
tests/llm/test_modelslab_llm.py	Pytest tests (7 cases)

Usage

LLM_PROVIDER=modelslab
LLM_NAME=meta-llama/Meta-Llama-3-8B-Instruct
MODELSLAB_API_KEY=your_key_here

Why ModelsLab?

• 300K+ developers using the API
• 10,000+ hosted open-source models on their own datacenter (vertically integrated with GPULab)
• OpenAI-compatible endpoint — zero friction to add
• Same integration pattern as the existing Novita provider in this repo

Checklist

• Follows existing OpenRouter/Novita provider pattern
• MODELSLAB_API_KEY wired through settings + model_utils provider_key_map
• Models added to registry with proper capabilities
• Pytest tests added covering instantiation, URL defaults, key fallback, and registration
• PEP 8 / ruff-compatible code

[vercel]

@adhikjoshi is attempting to deploy a commit to the Arc53 Team on Vercel.

A member of the Team first needs to authorize it.