Fact · Speech recognition
143seconds
In the faster-whisper GPU benchmark, openai/whisper took 2 minutes 23 seconds (143 seconds) with the large-v2 model in fp16 precision.
Checked against its source on October 17, 2026
In the source's own words
“| openai/whisper | fp16 | 5 | 2m23s | 4708MB |”
faster-whisper: README (GitHub, SYSTRAN), SYSTRAN, GitHub (SYSTRAN/faster-whisper), 2026, Large-v2 model on GPU table.
Open the original- What it measures
- The table row is for beam size 5. The README says the GPU runs used CUDA 12.4 on an NVIDIA RTX 3070 Ti 8GB.
- What it does not show
- A single run on one consumer GPU. The table also reports 4,708 MB of VRAM for this run.