Benchmark · June 2026
AI model benchmark
All models in one table - rating and key dimensions side by side. Subjective, but honest. Scroll sideways to see everything.
| Model | Rating | Best for | Context | Code | Images | Web / live data | Price |
|---|---|---|---|---|---|---|---|
ClaudeAnthropic
|
9.4 | Writing, code, analysis | very large | top | reads images | no | Free · Pro $20/mo |
ChatGPTOpenAI
|
9.2 | All-rounder | large | very good | yes | yes | Free · Plus $20/mo |
GeminiGoogle
|
8.8 | Google ecosystem | huge | good | yes | yes | Free · AI Pro |
GrokxAI
|
8.3 | Live data from X | large | good | yes | yes (X) | In X Premium+ |
DeepSeekDeepSeek
|
8.1 | Code, math, budget | large | very good | no | no | Free · cheap API |
PerplexityPerplexity
|
8.0 | Research with sources | medium | ok | no | yes (core) | Free · Pro $20/mo |
LlamaMeta
|
7.6 | Self-hosting | medium | good | some variants | no | Free (open) |
MistralMistral AI
|
7.3 | EU, privacy | medium | good | no | no | Freemium |
CopilotMicrosoft
|
7.0 | Microsoft 365 | medium | good | yes | yes | Freemium |
Rating 0-10 · approximate values, updated regularly