Dia
| Provider | Price | Provider model ID | Status | Media deployment details | Source |
|---|---|---|---|---|---|
 Segmind | Serverless Tariff not listedUsage $0.0072USD / gpu second Segmind · ServerlessUsage: $0.0072 Billed per GPU compute second, not generated media duration. Runtime varies. Checked 2026-10-03T14:45:20.473Z Price source ↗ | dia | Catalog-listedConditionsCatalog-listed. Account, region and endpoint restrictions may apply; check the source for deployment limits. | Not supplied | Official source ↗2026-09-28 |
Dialogue generation with optional reference-audio conditioning.
Media specifications
- Features
- Dialogue generation with optional reference-audio conditioning.
These describe the linked model reference. Individual provider endpoints can support different limits and options.
Specifications
- Catalog model ID
- nari-labs/dia
- Developer model ID / revision
- Not verified in cited sources
- Input modalities
- text, audio (optional reference)Developer repository ↗
- Output modalities
- audioDeveloper repository ↗
- Weight access
- Open weightsDeveloper repository ↗
- Weight license / access terms
- Apache 2.0Developer repository ↗
- Total parameters (billions)
- 1.6Developer repository ↗
- Download size (GB)
- Depends on checkpoint and precision; no verified file size
- Release date
- Not verified in cited sources
- Notes
- Specifications refer to the linked model documentation. Provider limits may differ. Unknown values are not inferred from the model name or listing date.
Catalog entries link to their specification and availability sources. This is a curated selection, not every model each provider offers. Best prices are the lowest collected USD tariffs, not a total-cost estimate. Token offers are selected by input price, with output from the same tariff; workload mix may change the cheapest provider. Listed variants may require batch processing, off-peak hours, or specific media settings. Price coverage varies by provider; collected tariffs include their source and observation time. Benchmark scores are a dated snapshot, not a live feed or a universal measure of quality. They apply to the shown evaluation configuration; listed prices may use different settings. Models without verified scores are not assumed to be weaker.