Users find Opus 5.5 has remarkably strong hearing, works via spectrograms
repligate · x · 2026-09-24
User @SkyeSharkie reports that Opus 5.5 demonstrates surprisingly strong audio perception: the model can create spectrograms to process audio tasks and even "listen" to its own work. The demo highlights a novel workflow where spectrograms serve as an intermediate representation for sound-related operations.
More from Models
- Gemini 4 rumored to be near release, possibly dropping next week — haider1 · 2026-09-24
- Open-weights CLM matches Jev's API surface, runs 4-13x faster with 75MB tunable heads — R_Duncan · 2026-09-24
- Jev open-source clones hit 25ms local, 50x faster than official; author runs 77-day trading backtest for $0.09 — anselm · 2026-09-24
- Putting a Decision Model Like JEV in Production: Thresholds Are Margins, Not Accuracy — sven_ai · 2026-09-24
- GPT-6 Astra agent plays Left 4 Dead 2 in first real-time run — imjustnewatai · 2026-09-24
- Same Prompt, Opposite Results: GPT-4 Goes Silent 30/30 Where GPT-3.5 Never Stops — rayanpal_ · 2026-09-24