Our take
Baseten is a US-based inference provider with strong compliance credentials and a curated catalogue of eight tracked offers. It emphasises security certifications, prompt privacy, and transparent throughput data over sheer breadth of model choice.
Use this for compliance-sensitive deployments that need an independent security attestation, a zero-retention option, and a verified commitment not to train on prompts. Pick it for throughput-predictable workloads where measured tokens-per-second enables reliable capacity planning. Skip it if you need a verified EU-region endpoint, or if your workload requires a deep catalogue of model choices.
- Strong compliance and privacy posture: independent security attestation, no training on prompts, and a zero-retention option — all three verified rather than claimed.
- Transparent throughput data for every offer: all eight include measured tokens per second, ranging from 31.5 to 190.5.
- Clear throughput-price tradeoff on at least one model: a higher-throughput tier of Z-AI GLM 5-2 delivers more than four times the speed for a proportionally higher price.
- Narrow catalogue depth versus open-weight specialists: eight tracked offers is the smallest catalogue among providers we track with comparable compliance data.
- No verified EU-region endpoint, limiting data-residency options.
- Premium pricing on several frontier models, with some of the most expensive tiers paired with relatively low throughput.
Privacy & data handling
- ✓
- Yes
- ✗
- No
- No answer on record
The colour says whether the answer favours you, not whether it is a yes — not training on your prompts earns a green cross, no zero-retention option earns a red one.
Prompt retention: none.