My goto web chat AI agents as of March 24, 2025

24 Mar, 2025

These are my current goto web chat AI agents:

• Grok 3 in X - base, Deepersearch, Think
• Qwen Max or QwQ with Search and Think, Search, Think, neither
• (for code, web prototypes, json responses) - Claude 3.7 Sonnet with extended thinking, web, code, artifacts
• ChatGPT o3-mini-high with or without web search
• Local models via open-webui/ollama on mac (you may pick smaller or larger variants depending on RAM on your mac):
  • SIGJNF/deepseek-r1-671b-1.58bit:latest
  • qwq:32b-fp16
  • gemma3:27b-it-fp16
  • r1-1776:70b-distill-llama-fp16
  • deepseek-r1:70b-llama-distill-fp16
  • phi4:14b-fp16
• For search and synthesis across research papers - ScholarQA from Allen AI (AI2)

Via API:

• openrouter models (has free models with rate limits)

Notes:

• Start with faster/cheaper options, try others only if they fail.
• I exclude anything exceeding about 20$ per month for web ui.
• I exclude local open models that have restrictive licenses.
• Some models require payment, some are free
• Each model has different policies around how they use your data. Most hosted models will likely use your data in one way or the other. Local models are generally private, if you have the hardware, and set it up properly.