Phying News
P非影 中文
Curated security research, vulnerabilities, advisories and tools for practitioners.

Why Model Routing Backfires and How to Build Agents That Don’t Burn Your Budget

jfrog.com | blog | #ai-agents | #model-routing | #llm-costs | #prompt-caching | #token-optimization | #jfrog-boost

Summary

JFrog Boost's co-founders explain why switching models mid-session backfires: prompt caches are model-specific, so the new model re-reads the whole conversation at full price, erasing routing savings.
Published
Collected

original ↗

Related coverage

blog · jfrog.com
Don’t Break the Agent: Lessons in Token Optimization
This coverage gives security teams current context for monitoring, validation, and remediation.
blog · jfrog.com
Boost Is Now In Public Preview
JFrog's Boost enters free public preview after saving 100 billion tokens across 1,000+ engineers; it trims the noise that makes input context roughly 70% of AI coding-agent spend.
blog · jfrog.com
Beyond Tokens SF: Best Ideas of the Evening
Recap of JFrog's Beyond Tokens SF event: Fly persists release context — commits, PRs, decision records — so agents stop losing intent, Boost cuts token waste, and NanoClaw sets security boundaries.
blog · about.gitlab.com
How GitLab reduced code-per-agentic-flow ratio by 45%
GitLab's Flow Registry compiles declarative YAML into LangGraph flows for Duo Agent Platform, cutting code per agentic flow by 45% while improving reliability and maintainability.
blog · about.gitlab.com
New MCP tools help platform teams scale automation safely
GitLab 19.4 expands the GitLab MCP server tool set so agents can run pipelines, open merge requests, and triage vulnerabilities, while writes, merges, and deletes still wait for a human reviewer.
blog · about.gitlab.com
GitLab Duo CLI takes a task from goal to done
GitLab 19.4 adds a /goal slash command to GitLab Duo CLI, letting developers delegate open-ended tasks that run to completion with verification, instead of re-prompting after every chat turn.
back