IBM

Latest AI news, models and releases from IBM. ['ChartNet', 'Granite 4.0', 'Granite-Docling-258M', 'Granite Speech', 'Granite Speech 4.0 1B', 'Granite Speech 4.1 2B', 'Granite Speech 4.1 2B-NAR', 'Granite-Vision-3.3-2B', 'IBM gender classifier', 'сверхпроводящий квантовый компьютер IBM']

Official site ↗

Research 🇺🇸

FFASR Leaderboard Launched: Benchmark for Evaluating ASR in Real Acoustic Conditions

Hugging Face and Treble Technologies have launched the Far-Field ASR (FFASR) Leaderboard, the first open community-driven benchmark for evaluating ASR models under realistic far-field conditions like reverberation, background noise, and microphone distance. Initial results show far-field WER at low SNR is several times higher than near-field WER, highlighting a significant gap between benchmark performance and real-world deployment.

Treble Technologies

Hugging Face

OpenAI

IBM

Cohere

ScarfBench: A Benchmark for Evaluating AI Agents in Enterprise Java Framework Migration

ScarfBench is an open benchmark for evaluating AI agents on cross-framework migration tasks in Enterprise Java, covering Spring, Jakarta EE, and Quarkus. It tests whether migrated applications build, deploy, and preserve behavior. Current agents show overconfidence and struggle with environment issues, with configuration dominating migration effort.

IBM

Anthropic

Hugging Face blog27.07 · 09:05

Show more ↓