Probed: missed a spoken codeword inside the clip. Any comment it makes on sound is invented.
SHAPE: yellow square AUDIO: NONE
Vision-only in practice — judge the frames, it cannot hear the clip.
Probed deaf to audio inside a video clip (2026-08-08): it named the yellow square in the probe clip correctly, but returned NONE for a codeword spoken aloud in that same file. OpenRouter billed 0 audio tokens for the call.
So anything GLM-5V Turbo says about music, voices, mix or sound design in a clip is invented — it does not volunteer that it heard nothing. Judge it on the frames only.
GLM-5V-Turbo is Z.ai’s first native multimodal agent foundation model, built for vision-based coding and agent-driven tasks. It natively handles image, video, and text inputs, excels at long-horizon planning, complex coding,...
| Field | Per token (USD) | Per million |
|---|---|---|
| prompt | 0.0000012 | $1.20 |
| completion | 0.000004 | $4.00 |
| input_cache_read | 0.00000024 | $0.240 |
{
"temperature": 1,
"top_p": 0.95,
"top_k": null,
"frequency_penalty": null,
"presence_penalty": null,
"repetition_penalty": null
}7 answersin this database. A round opens every model's answer to the same prompt.