Docs: remove max tokens. (#6198)

### What problem does this PR solve?

#6178

### Type of change

- [x] Documentation Update

docs/references/http_api_reference.md

         "language": "English",
         "llm": {
             "frequency_penalty": 0.7,
-            "max_tokens": 512,
             "model_name": "qwen-plus@Tongyi-Qianwen",
             "presence_penalty": 0.4,
             "temperature": 0.1,
             "language": "English",
             "llm": {
                 "frequency_penalty": 0.7,
-                "max_tokens": 512,
                 "model_name": "qwen-plus@Tongyi-Qianwen",
                 "presence_penalty": 0.4,
                 "temperature": 0.1,
                             "cite": true,
                             "frequency_penalty": 0.7,
                             "llm_id": "gpt-4o___OpenAI-API@OpenAI-API-Compatible",
-                            "max_tokens": 256,
                             "message_history_window_size": 12,
                             "parameters": [],
                             "presence_penalty": 0.4,
                                 "frequency_penalty": 0.7,
                                 "llm_id": "gpt-4o___OpenAI-API@OpenAI-API-Compatible",
                                 "maxTokensEnabled": true,
-                                "max_tokens": 256,
                                 "message_history_window_size": 12,
                                 "parameters": [],
                                 "presencePenaltyEnabled": true,