HELM TORR - Tab Fact - Robustness Leaderboard · The Aggregate

HELM TORR - Tab Fact - Robustness: leaderboard

Metric: Robustness (0-100). Source: crfm.stanford.edu. 14 models tracked.

Top models

#	Model	Score
1	Claude 3.5 Sonnet (20241022)	67
2	GPT-4o (2024-11-20)	57
3	Gemini 1.5 Pro (002)	55
4	DeepSeek V3	50
5	Llama 3.1 405B Instruct	47
6	Gemini 1.5 Flash (002)	43
7	Qwen 2 72B Instruct	40
8	Claude 3.5 Haiku (20241022)	36
9	Llama 3.1 70B Instruct	26
10	GPT-4o Mini (2024-07-18)	26
11	Mistral 7B Instruct (v0.3)	21
12	Llama 3.1 8B Instruct	2

Interactive version: theaggregate.ai/benchmark?slug=helm-torr-tab-fact-robustness · How It Works · Data refreshed daily, snapshot 2026-09-19.