Speech to Text now accepts vad_threshold to tune the voice-activity gate that skips non-speech audio.

The useful part is control over quieter or noisier speech instead of treating the gate as fixed.

It sits in streaming and batch Speech to Text, giving builders a direct setting for narrowband telephony audio.

• Jul 23, 2026OpenAI pricing diff removes Business and Enterprise plan details.
• Jul 23, 2026Anthropic model card now lists Claude Fable 5 and Mythos 5 availability.
• Jul 23, 2026Anthropic pricing removed its team plan recommendation quiz from the monitored page.

This is the world we build in. It moves this fast. We move with it.