New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
TriviaQA benchmark
AI model leaderboard for the TriviaQA benchmark. Compare how large language models score on TriviaQA, see the full ranking, and understand what this AI benchmark measures. OLMo 2 32B currently leads with 87.3. Closed-book question answering benchmark of trivia questions sourced from quiz websites.
Leaderboard
| # | Model | Organization | Score | Variant | Source |
|---|---|---|---|---|---|
| #1 | OLMo 2 32B | Allen Institute for AI | 87.3 | base-from-hf-readme | official ↗ |
| #2 | GPT-5 | OpenAI | 86.32 | thinking-high-cited-ring-1t | official ↗ |
| #3 | gemini-2.5-pro | Google DeepMind | 84.45 | cited-ring-1t | official ↗ |
| #4 | DeepSeek-V3.1-Terminus | DeepSeek | 82.77 | thinking-cited-ring-1t | official ↗ |
| #5 | Gemma 2 9B | Google DeepMind | 81.8 | base-cited-olmo2 | official ↗ |
| #6 | Llama 2-13B | Meta AI | 81.3 | base-cited-olmo2 | official ↗ |
| #7 | Llama 3.1-8B | Meta AI | 80.3 | base-cited-olmo2 | official ↗ |
| #8 | Qwen3-235B-A22B-Thinking (Jul 2025) | Qwen | 79.63 | thinking-cited-ring-1t | official ↗ |
| #9 | Ring-1T | Ant Group | 78.59 | thinking | official ↗ |
| #10 | Qwen2.5-7B | Qwen | 69.4 | base-cited-olmo2 | official ↗ |
| #11 | DCLM 7B | Apple | 52.7 | from-hf-readme | official ↗ |
Frequently asked questions about TriviaQA
What is the TriviaQA benchmark?
Closed-book question answering benchmark of trivia questions sourced from quiz websites.
How is the TriviaQA benchmark scored?
TriviaQA is scored using the exact_match metric, where a higher score is better. GenAIList aggregates reported scores from model providers and papers into a single ranked leaderboard.
Which AI model scores highest on TriviaQA?
As of the latest reported scores on GenAIList, OLMo 2 32B achieves the highest result on TriviaQA with a score of 87.3.
Is a higher TriviaQA score better?
Yes. On TriviaQA a higher score indicates better performance, so models near the top of the leaderboard are the strongest.