PostHog Shares How It Uses Jev to Gate Expensive LLM Calls
PostHog has published a thread on how it uses Jev decision models in production, from privacy scrubbing to chart picking, and where they fall short.
Integrations, competing decision APIs, open-weight decision models and developer infrastructure.
PostHog has published a thread on how it uses Jev decision models in production, from privacy scrubbing to chart picking, and where they fall short.
With Google’s Gemini Decision 1 joining OpenAI and Microsoft at $0.10 or less per million tokens, decision models have surged to one in four Vercel AI Gateway requests.
Jev leads OpenRouter’s decision-model traffic with over 5.8 billion requests in a single month, even as rivals like Cloudflare, OpenAI, and Sarvam replicate the pattern fast.
Hugging Face added native DecisionPipeline, AutoModelForDecision, and DecisionConfig APIs to Transformers so open decision models share one inference contract.
Pydantic AI introduced a unified DecisionModel abstraction mapping Python output types directly to Jev, System One servers, OpenAI Decisions, and Gemini Decision 1.
Forbes examines how Perplexity’s $0.02/M open-weight Pplx-Decider and Liquid AI’s $0.04/M D1 are turning TypeSafe’s new category into a fierce price and architecture battle.
On October 6, OpenAI moved its Decisions API into public beta on GPT-6 Luna on the same day Perplexity open-sourced its 27B Pplx-Decider under Apache 2.0.
Two weeks after Jev launched, OpenAI unveiled a fast-tracked Decisions API preview at DevDay 2026 on the same day PostHog released Jeeves, an open 9B reasoning decider.
Within just over a week of TypeSafe’s launch, open-source builders released Laya, Nimble, Kev, SemIf, Rizzo Flow, Von, and NanoJev—all rallying around Jev’s typed request contract.
Developer infrastructure moved immediately after launch: LangChain shipped guardrail middleware and LangSmith eval guides while Vercel’s AI Gateway saw 13% paid-team trial on day one.