Model comparison
Editorial comparison, not marketing. Scores on a 10-point scale. Updated as models and versions change.
Last updated: 5/17/2026. Don't use this table as a final verdict — always test for your own needs.
| Model | Reasoning | Coding | Writing | Research | Multimodal | Long context | Tool use |
|---|
ChatGPT (GPT-5 class) OpenAI | | | | | | | |
Claude Anthropic | | | | | | | |
DeepSeek DeepSeek | | | | | | | |
Gemini Google | | | | | | | |
Grok xAI | | | | | | | |
Local LLM (Llama / Qwen / DeepSeek через Ollama) Local | | | | | | | |
Mistral Mistral | | | | | | | |
Perplexity Perplexity | | | | | | | |
Qwen Alibaba | | | | | | | |
ChatGPT (GPT-5 class)
OpenAI · updated 5/17/2026
Strengths
- Versatility
- Ecosystem (plugins, custom GPTs)
- Voice and multimodal
- Good quality
Weaknesses
- Shorter context than Claude/Gemini
- A paid subscription is needed for the best model
Best use cases
- A universal assistant
- Text, ideas, brainstorming
- Voice input and working with images
Pricing: Free with limits; ChatGPT Plus / Pro are paid
Privacy: You can disable training on your data in the settings
Claude
Anthropic · updated 5/17/2026
Strengths
- Long context
- Careful work with documents
- Fewer confident errors
- Claude Code for development
Weaknesses
- Fewer plugins
- Sometimes overly cautious
Best use cases
- Long documents
- Writing and analysis
- Development via Claude Code
Pricing: Free; Claude Pro / Max are paid
Privacy: By default doesn't train on API/Pro user data