AI Paper Atlas

Research map / Benchmarks / AQUA-RAT

AQUA-RAT leaderboard

10 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on AQUA-RAT. Best accuracy: 86.2 (HPR@5 (32B hinter), AAAI 2026).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

AQUA-RAT · accuracy

10 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
186.2HPR@5 (32B hinter)
Qwen2.5-7B-Instruct (practitioner)
Efficient Thought Space Exploration Through Strategic Intervention
AAAI 2026 · Table 2
Reasoning and chain of thought
277.56PRESTO
GPT-4.1
PRESTO: Preimage-Informed Instruction Optimization for Prompting Black-Box LLMs
NeurIPS 2025 · Table: Chain-of-Thought Prompt
Language, safety and interpretability
373.62SelfOrg
Qwen-2.5-3B-Instruct
Stochastic Self-Organization in Multi-Agent Systems
ICLR 2026 · text
LLM agents
464.6Promptbreeder few-shot
PaLM 2-L
Promptbreeder: Self-Referential Self-Improvement via Prompt Evolution
ICML 2024 · Table 1
Reasoning and chain of thought
554.724INSTINCT
ChatGPT
Use Your INSTINCT: INSTruction optimization for LLMs usIng Neural bandits Coupled with Transformers
ICML 2024 · Table: zero-shot CoT instructi
Language, safety and interpretability
654.724ZOPOLocalized Zeroth-Order Prompt Optimization
NeurIPS 2024 · Table: CoT prompt results
Language, safety and interpretability
747.2MAmmoTH-Coder-7B
Code-Llama
MAmmoTH: Building Math Generalist Models through Hybrid Instruction Tuning
ICLR 2024 · Table 3
Reasoning and chain of thought
841.73MuToR
Gemma 2B
Multi-Token Prediction Needs Registers
NeurIPS 2025 · Mathematical reasoning table
Instruction and fine-tuning
933.5Zero-shot-CoT
text-davinci-002
Large Language Models are Zero-Shot Reasoners
NeurIPS 2022 · Table: Accuracy comparison of
Language, safety and interpretability
1030.16MIO
Qwen2.5-7B-Instruct
The Hidden Link between RLHF and Contrastive Learning
ICML 2026 · Table 2
Alignment and preferences

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.