Large language model
Verdict
Qwen3 30B A3B Thinking 2507 from Alibaba. Open weights (Apache 2.0). 81K context. $0.2/$2.4 per 1M tokens.
Training data policy: N/A — open weights. Self-host for full control.
DPDP Act (India): Self-host in India for full DPDP compliance.
Data residency: Self-hosted
Self-host in India