A simple question on Hacker News has sparked a broader conversation about how developers choose between the growing number of large language models available today. A user asked the community whether any curated list exists ranking the best LLMs by use case.
The fragmented LLM market
The question reflects a real challenge facing developers: with dozens of capable models now available from companies like OpenAI, Anthropic, Google, Meta, and Mistral, choosing the right one for a specific task has become a research project in itself. Each model has different strengths, pricing structures, and context window limits.
The original poster, daytonflyers, received two comments within an hour but no comprehensive list was provided in the thread. The question remained largely unanswered in the discussion, which itself underscores the difficulty of the problem. There is no widely accepted standard for categorizing LLM performance by use case.
Developers often rely on benchmarks like MMLU for general knowledge, HumanEval for coding, or MT-Bench for conversational quality, but these scores don't always translate to real-world performance for specific workflows.
How developers actually choose
In practice, most developers pick models based on a combination of factors: task type, cost, latency requirements, and integration compatibility. Code generation tasks favor models like Claude or GPT-4, while long-document analysis might benefit from models with larger context windows. Creative writing tasks often work better with models tuned for style and coherence.
The lack of a definitive ranked list means developers often test multiple models against their specific use case before committing. Tools like Chatbot Arena, which pits models against each other in blind comparisons, have gained traction as a way to evaluate relative strengths without relying solely on synthetic benchmarks.
As the number of available models continues to grow, the need for practical, use-case-specific comparisons becomes more pressing. The Hacker News thread, though brief, highlights a gap in the current ecosystem that no single resource has yet filled.