Share your thoughts, 1 month free Claude Pro on usSee more

Sentence Stress Reasoning on StressPresso

89.6SSR Accuracy

Human Annotators

Updated 11d ago

Evaluation Results

Method	Links
Human Annotators 2025.05		89.6	96	-
StresSLM 2025.05		87.6	-	-
gpt-4o 2025.05		83.6	-	-
Llama-3.1-8B-Instruct 2025.05		81.1	-	-
gpt-4o-mini 2025.05		80.1	-	-
WhiStress → gpt-4o 2025.05		79.7	-	-
WhiStress → Llama-3.1-8B-Instruct 2025.05		76.2	-	-
Qwen2-7B-Instruct 2025.05		74.2	-	-
WhiStress → gpt-4o-mini 2025.05		73.2	-	-
Gemini-2.5-Pro 2025.05		72.7	-	-
WhiStress → Qwen2-7B-Instruct 2025.05		69.3	-	-
Qwen-7B-Chat 2025.05		64.8	-	-
gpt-4o-audio 2025.05		64.8	-	-
Qwen3-Omni-30B-Instruct 2025.05		64.8	-	-
WhiStress → Qwen-7B-Chat 2025.05		62.3	-	-
Audio-Flamingo-3 2025.05		52.9	-	-
LLaMA-Omni 2025.05		52.9	-	-
SALMONN 2025.05		52.4	-	-
Qwen2Audio-7B-Instruct 2025.05		51.4	-	-
Phi-4-multimodal-instruct 2025.05		50.4	-	-
Gemini-2.5-Pro 2025.05		-	-	3.29
gpt-4o-audio 2025.05		-	-	2.84
Audio-Flamingo-3 2025.05		-	-	2.39
Qwen3-Omni-30B-Instruct 2025.05		-	-	2.42
Qwen2Audio-7B-Instruct 2025.05		-	-	2.79
SALMONN 2025.05		-	-	2.19
LLaMA-Omni 2025.05		-	-	1.92
Phi-4-multimodal-instruct 2025.05		-	-	2.73
StresSLM 2025.05		-	-	3.83