Unverified50% confidenceFactExact time
Evaluation results on current frontier MLLMs show that the models perform strongly on local perception tasks but are clearly weak in long-horizon audio-visual reasoning.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
UnverifiedAlthough full-modal large models support voice input and output, their scores on agent benchmarks are far lower than those of mainstream reasoning models.70% similarUnverifiedThe average evaluation metrics of a new LLM may be better than those of an older model, but regressions can occur in niche critical scenarios.69% similarUnverified非自回归语音模型通常比自回归模型快10-100倍，但表现力受限于预测器准确性67% similarUnverifiedWhen using smaller or quantized models in local deployments, the probability of AI hallucinations may be higher than with large commercial models.66% similarUnverifiedExisting model evaluation frameworks (such as HumanEval, MBPP, and various Arena benchmarks) primarily measure capability ceilings and rarely incorporate end-to-end task completion rate as an evaluation dimension.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1012774API
curl https://kongchang.com/api/v1/knowledge/claims/1012774MCP
get_claim(id=1012774)