Add CheaperInferenceLLM provider

[aiapienthusiast]

Adds CheaperInferenceLLM, a model class for Cheaper Inference.
It follows the same pattern as NovitaLLM (#268).

Cheaper Inference is an OpenAI-compatible LLM gateway.
One API key gives access to models from several labs.

The class subclasses OpenAILLM and only sets the client and the base URL.
Like NovitaLLM, it records token counts and sets the cost to 0, because LiteLLM has no prices for these models.

Usage:

from evoagentx.models import CheaperInferenceConfig, CheaperInferenceLLM

config = CheaperInferenceConfig(model="gpt-5.4-mini", cheaperinference_key="your-key")
llm = CheaperInferenceLLM(config=config)

Details

Files:

• evoagentx/models/cheaperinference_model.py: new CheaperInferenceLLM (alias cheaperinference), base URL https://api.cheaperinference.com/v1.
• evoagentx/models/model_configs.py: new CheaperInferenceConfig.
• evoagentx/models/__init__.py: export.
• docs/modules/llm.md, docs/zh/modules/llm.md: usage section.

Tests:

• ruff check evoagentx: pass.
• pytest tests/src/models: 36 passed.

Live check: gpt-5.4-mini returned a reply through CheaperInferenceLLM.generate() and async_generate().