feat(streaming): add universal-3-6-pro speech model

[zkleb-aai]

Summary

Adds universal-3-6-pro to the streaming SpeechModel enum so it can be selected via StreamingParameters:

params = StreamingParameters(
    sample_rate=16000,
    speech_model=SpeechModel.universal_3_6_pro,  # or "universal-3-6-pro"
)

The client passes speech_model straight through to the speech_model query parameter with no allowlist validation, so the enum member is the only change needed. It applies to both RealTimeTranscriber and AsyncRealTimeTranscriber, which share these models.

Scope is streaming only — the pre-recorded SpeechModel and the sync SyncSpeechModel are untouched.

Changes

• assemblyai/streaming/v3/models.py — add universal_3_6_pro to SpeechModel
• tests/unit/test_streaming.py — assert the model reaches the connection URL, mirroring the existing universal-3-5-pro test
• assemblyai/__version__.py — 1.1.0

Testing

• pytest tests/unit — 469 passed, 0 failed
• ruff check surfaces 10 pre-existing I001 import-sort errors in tests/unit/test_*.py files this PR doesn't touch; nothing new
• ruff format --check clean on the changed files

🤖 Generated with Claude Code