ArisGlobal reverses prompt order on Amazon Bedrock to cache shared prefix
Start here
ArisGlobal changed how its NavaX AI engine orders prompts on Amazon Bedrock, caching the shared prompt prefix instead of each source document, which AWS says reduced effective input cost and let the company add tenants without proportional cost growth. The shared prefix grew from about 72,000 to 180,000 tokens in six months, pushing per-case inference costs up more than 200%. NavaX has processed over 1 million real transactions in regulated pharmaceutical environments since its first Bedrock deployment with Anthropic's Claude in March 2025. Prompt caching on Bedrock has a default TTL of 5 minutes; select models support a 1-hour TTL via "ttl": "1h". For Claude Sonnet 4.5 the minimum cacheable prompt is 1,024 tokens, and model availability varies by Region.
Generated from reports · updated 8 h ago
Report timeline
Follow the reports to see the event from different sides.
- AWS Machine Learning BlogHow ArisGlobal scaled generative AI inference on Amazon Bedrock
ArisGlobal reversed the prompt order in its NavaX AI engine on Amazon Bedrock, caching the shared prompt prefix instead of each source document to cut effective input costs. The shared prefix grew from about 72,000 to 180,000 tokens in 6 months, pushing per-case inference costs up more than 200%. NavaX has processed over 1 million real transactions in regulated pharmaceutical environments since its first Amazon Bedrock deployment with Anthropic's Claude in March 2025.
Heat for this event
- Current, comparable range
- 8
- Peak, comparable range
- 10November 18 01:00
- Change in 24 hours
- –
The trend compares the same participants over hours observed continuously, so its range can be narrower than the current heat count. Hover or tap the chart for hourly heat; the arrow keys move through it.