New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
Compare AI models head-to-head
Build your own LLM comparison: pick the models you care about and get a dedicated, shareable results page with their specs and benchmark scores side by side.
Choose models
Models (0 selected)
Benchmarks (auto)
Popular models compared
| Model | GPQA Diamond | MMLU-Pro | MMLU | AIME 2025 | IFEval | HumanEval |
|---|---|---|---|---|---|---|
| Kimi K2 | 74.2% | 81.9% | 89.5% | 75.2% | 90.99% | 87.1% |
| DeepSeek-R1-Distill-Qwen-7B | 49.1% | 53.5% | 50.72% | 56.7% | 60.5% | 79.89% |
| DeepSeek-R1-Distill-Qwen-7B | 49.1% | 53.5% | 50.72% | 56.7% | 60.5% | 79.89% |
| DeepSeek-V3 | 59.1% | 76% | 88.5% | 28.8% | 88% | 94.6% |
| Claude 3.5 Sonnet | 65% | 78% | 88.7% | 7.4% | 90.1% | 93.7% |
| Granite-4.0-H-Micro | 32.15% | 43.48% | 67.43% | — | 86.94% | 81% |
| Granite-4.0-H-Tiny | 32.59% | 44.94% | 68.65% | — | 84.78% | 83% |
| Granite-4.0-H-Small | 40.63% | 55.47% | 78.44% | — | 89.87% | 88% |
A teaser of the most-benchmarked models. Pick models above to generate a full comparison page.
How to compare AI models
Comparing large language models is rarely about a single number. A model that leads on a coding benchmark such as SWE-bench may trail on math or reasoning. Pick any models above to get a dedicated comparison page with their specs and scores side by side across coding, reasoning, math, agentic and multimodal benchmarks.
Browse the full AI benchmarks leaderboard, explore code models or language models, or open any model page to see its complete scorecard and pricing.
Frequently asked questions
How do I compare AI models on GenAIList? ▶
Pick one or more models and the benchmarks you care about, then build a side-by-side matrix of their scores. Each score links back to the model and benchmark pages, and is sourced from the original system cards and technical reports.
Which AI model is the best? ▶
There is no single best AI model — it depends on the task. A model that tops a coding benchmark like SWE-bench may rank lower on math (AIME) or reasoning (GPQA). Use the comparison matrix to weigh the benchmarks that matter for your use case.
Is ChatGPT or Claude better? ▶
It depends on the workload. Compare the latest OpenAI and Anthropic models head-to-head across coding, reasoning and math benchmarks to see where each one leads.
Where do the benchmark scores come from? ▶
Scores are aggregated from official model cards, technical reports and standard public evaluations, and are updated as new results are published.