*Fabulous* LLM Tools
Updated: 2025-08-19Here I list all sorts of tools that use LLMs (Large Language Models). I feel ambivalent about hosting a ressource list that use those insidious tools. When I get around to it, I will link a blog post that talks about the problems inherent to LLMs and other «AI» tools.
In the mean time, here is a non-exhaustive list of articles on the subject of LLM assisted creation (mostly coding).
- Here’s how I use LLMs to help me write code
- A cheat sheet for why using ChatGPT is not bad for the environment
- A trick to feel less like cheating when you use LLMs
- My LLM codegen workflow atm | Harper Reed's Blog
- programming with llms- crawshaw - 2025-01-06
- Everything I've learned so far about running local LLMs
- An example of LLM prompting for programming
- How I write code using Cursor: A review
VERY IMPORTANT → Hallucinations in code are the least dangerous form of LLM mistakes
Providers
To check prices for different models and providers, you can refer to this incredible LLM Pricinc Calculator. It has the pricing for Input and Output costs for different models, as well as a calculator that takes as parameters the number of input and output tokens. This tool is made and maintained by the amazing Simon Willison.
Cloud
Openrouter.ai is a platform that enables you to have a single api key to be used with multiple models, including private ones.
Chats
Most, if not all services listed in the Chat section, will use your data for training purposes. They are not suitable for privacy sensitive queries.
-
Claude, by Anthropic - $
-
Chatgpt by OpenAI - $
-
Gemini, by Google - $
-
AI Studio, also by Google.
-
Le Chat, by Mistral ($)
-
Deepseek, by Deepseek $
-
Qwen, by QwenAI $
-
GLM, by z.ai
-
Grok, by x.ai → use is ethical very questionable, since it is built by a E. Musk company. Know for being extreme right oriented, and will actively provide erroneous opinions as facts.
-
Phind, by Phind. Creates a lot of diagrams
-
Perplexity, by PerplexityAI ($)
-
Poe, access to multiple models. #freetier
-
Meta ($)
-
Amazon $
-
LMArena, free random models
Local
- Llama.ccp, inference engine in C/C++, very efficient and is the basis for most other serving tools.
- Ollama, go based hosting tool, with a good DL directory of models. It serves an OpenAI compatible API endpoint.
- Llamafile cross-platform model in a file, that runs both with CLI interface, web interface, and can also serve as an OpenAI compatible API endpoint.
- LM Studio, electron based local interface, model curated, local server capable app. Not entirely open source though.
CLI
- llm, a python CLI tool with a really cool range of features and uses, also by Simon Willison.
- Aider.chat, a python cli llm «agent» that sort of automates programming. You can connect all sorts of providers to it, including your own locally served models (via Ollama,Llama.ccp, VLLM, LM Studio).
- Anthropic Code
- OpenAi Codex