AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

BFCL Audio: An Audio Function Calling Evaluation for Large Language Models

Huanzhi Mao, Aditya Ghai, Imra Dawoodani, Tony A Ginart, Shishir G Patil, John Emmons, Joseph E. Gonzalez

ICML 2026 · Large language models › Language, safety and interpretability · approach: Speech and audio language models

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add background noise to speech input (ablation)GPT-4o-audio · BFCL True Audio overalloverall accuracy: 60.4 -> 50.9Hurt
add background noise to speech input (ablation)Qwen3-Omni-Flash · BFCL True Audio overalloverall accuracy: 55.8 -> 48.2Hurt
add background noise to speech input (ablation)Gemini-2.5-Pro · BFCL True Audio overalloverall accuracy: 51.1 -> 45.4Hurt
add competing speech (double-talk) (ablation)clean utterances · intent preservationintent preservation: 42.4%Hurt
add device artifacts (mic rubbing, mumbling, fades, gain changes) (ablation)clean utterances · intent preservationintent preservation: 98.8% (mic rubbing), 100.0% (mumbling)No clear effect
inject discourse-altering disfluencies (self-corrections, ellipsis, restarts) (ablation)ASR->LLM pipeline · BFCL Text Audiodownstream failuresHurt
inject surface-level disfluencies (fillers, repetitions, contractions) (ablation)ASR->LLM pipeline · BFCL Text Audiodownstream failuresNo clear effect
verbalize file paths in speech (ablation)end-to-end audio models · BFCL True Audio (clean)errorsHurt
use end-to-end audio input instead of ASR transcripts (clean audio)GPT-4o-audio · BFCL Audio overalloverall accuracy: 58.6 -> 60.4Improved
use end-to-end audio input instead of ASR transcripts (noisy audio)GPT-4o-audio · BFCL Audio overalloverall accuracy: 58.6 -> 50.9Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy