TriviaQA

Models scored

115

Best score

0.88

Metric

EM

TriviaQA is an AI evaluation tracked by Epoch AI, with 115 scored model versions on record. The scores here are reported by third parties rather than produced by Epoch AI running the evaluation itself.

Scored models span December 2021 to December 2024. The best score recorded is 0.88, measured as EM. Models come from Anthropic, OpenAI, Meta AI, Google Research, Google DeepMind, Mistral AI, DeepSeek, Alibaba and others.

Full record

Score metric: EM
Models scored: 115
Best score recorded: 0.88
Earliest model scored: Dec 8, 2021
Most recent model scored: Dec 26, 2024
Third-party reported: Yes

Leaderboard

01	Llama-2-70b-hf	Meta AI	0.88
02	claude-2.0	Anthropic	0.88
03	claude-1.3	Anthropic	0.87
04	PaLM 2-L		0.86
05	LLaMA-65B	Meta AI	0.86
06	gpt-3.5-turbo-1106	OpenAI	0.86
07	gpt-4-0613	OpenAI	0.85
08	Llama-2-34b	Meta AI	0.85
09	LLaMA-33B	Meta AI	0.84
10	DeepSeek-V3	DeepSeek	0.83
11	Llama-3.1-405B	Meta AI	0.83
12	Mixtral-8x7B-v0.1	Mistral AI	0.82
13	PaLM 2-M		0.82
14	PaLM 540B	Google Research	0.81
15	DeepSeek-V2	DeepSeek	0.8
16	falcon-40b	Technology Innovation Institute	0.8
17	Llama-2-13b	Meta AI	0.8
18	claude-instant-1.1	Anthropic	0.79
19	claude-instant-1.2	Anthropic	0.79
20	LLaMA-13B	Meta AI	0.78
21	GLaM (MoE)	Google	0.76
22	PaLM 2-S		0.75
23	Mistral-7B-v0.1	Mistral AI	0.75
24	Phi-3-medium-128k-instruct	Microsoft	0.74
25	Llama-2-7b	Meta AI	0.74
26	mpt-30b	MosaicML	0.74
27	gemma-7b	Google DeepMind	0.72
28	Qwen2.5-72B	Alibaba	0.72

More benchmarks

Epoch Capabilities Index MMLU Terminal-Bench GPQA Diamond OTIS Mock AIME 2024-2025 GSM8K ARC-AGI ARC-AGI-2

SourceEpoch AI, 'AI Benchmarking Hub'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/benchmarks. Licensed under CC BY 4.0.

← All benchmarks