Sophon-Realtime Models by Adventists AI
Adventists AI releases · 6 in the last 12 monthsChangelog →
6 models2026
Last refreshed 2026-11-19. Next refresh: weekly.
Details
ResearcherAdventists AI
LicenseApache 2.0OSI-approved
Commercial usePermitted
Models6
Released2026
Links
WebsiteAbout
Adventists AI’s full-duplex spoken-dialogue models.
Decision facts
- Best fit
- speech to speech
- Capability starting point
- Sophon-Realtime 4B E2E
- Lowest tracked input
- Not tracked
- Closest related family
- 360Zhinao
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
6 in view
Sophon-Realtime 4B TagCurrent
Use when the workload needs speech to speech, audio, and open source.
2026-11speech to speechaudioopen source
Sophon-Realtime 4B FlashCurrent
Use when the workload needs speech to speech, audio, and open source.
2026-11speech to speechaudioopen source
Sophon-Realtime 4B E2ECurrent
Use when the workload needs speech to speech, audio, and open source.
2026-11speech to speechaudioopen source
Sophon-Realtime 32B TagCurrent
Use when the workload needs speech to speech, audio, and open source.
2026-11speech to speechaudioopen source
Sophon-Realtime 32B FlashCurrent
Use when the workload needs speech to speech, audio, and open source.
2026-11speech to speechaudioopen source
Sophon-Realtime 32B E2ECurrent
Use when the workload needs speech to speech, audio, and open source.
2026-11speech to speechaudioopen source
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Sophon-Realtime 4B Tag | Use when the workload needs speech to speech, audio, and open source. | 2026-11 | speech to speechaudioopen source | Current |
| Sophon-Realtime 4B Flash | Use when the workload needs speech to speech, audio, and open source. | 2026-11 | speech to speechaudioopen source | Current |
| Sophon-Realtime 4B E2E | Use when the workload needs speech to speech, audio, and open source. | 2026-11 | speech to speechaudioopen source | Current |
| Sophon-Realtime 32B Tag | Use when the workload needs speech to speech, audio, and open source. | 2026-11 | speech to speechaudioopen source | Current |
| Sophon-Realtime 32B Flash | Use when the workload needs speech to speech, audio, and open source. | 2026-11 | speech to speechaudioopen source | Current |
| Sophon-Realtime 32B E2E | Use when the workload needs speech to speech, audio, and open source. | 2026-11 | speech to speechaudioopen source | Current |
Release Timeline
1 release group2026-11
6 current
Sophon-Realtime 32B E2E
Currentspeech to speechaudioopen source
Sophon-Realtime 32B Flash
Currentspeech to speechaudioopen source
Sophon-Realtime 32B Tag
Currentspeech to speechaudioopen source
Sophon-Realtime 4B E2E
Currentspeech to speechaudioopen source
Sophon-Realtime 4B Flash
Currentspeech to speechaudioopen source
Sophon-Realtime 4B Tag
Currentspeech to speechaudioopen source
Specifications(6 models)
| Model | Released |
|---|---|
| Sophon-Realtime 4B Tag | 2026-11 |
| Sophon-Realtime 4B Flash | 2026-11 |
| Sophon-Realtime 4B E2E | 2026-11 |
| Sophon-Realtime 32B Tag | 2026-11 |
| Sophon-Realtime 32B Flash | 2026-11 |
| Sophon-Realtime 32B E2E | 2026-11 |