Exploring LLM costs

PostHog attaches per-call cost metadata to every $aigeneration and $aiembedding event at ingestion time. Every cost question reduces to an aggregation over those two event types — the interesting variation is only in how you group, filter, and compare.

This skill covers the common cost investigations: total spend, breakdowns (model, provider, user, trace, custom property), token and cache-hit analysis, regression debugging, and materializing results as insights, dashboards, or alerts.

Tools

Core rules

Three rules cover most of what goes wrong:

• Sum $aitotalcost_usd for rollups, never the components. Components drop

request and web-search fees. The UI's cost cells sum $aitotalcostusd over event IN ('$aigeneration', '$ai_embedding'); mirror that. Full schema and rationale in [cost properties](./references/cost-properties.md).

• Always include both $aigeneration and $aiembedding in cost queries

unless the project demonstrably does not use embeddings — missing them silently under-counts. $aitrace and $aispan carry no rollup cost; some SDK wrappers duplicate $aitotalcostusd onto $aitrace so don't include it in rollups or you'll double-count.

• Always set a time range. Cost queries without one scan the full events table.

$aitotalcostusd is set at ingestion via one of three paths (passthrough, custom pricing, automatic lookup). When a cost looks wrong, read $aicostmodelsource first — see [cost sources](./references/cost-sources.md) for the precedence rules and a diagnostic query.

Cache-hit math depends on whether the provider reports cache tokens inclusively or exclusively of $aiinputtokens. Always branch on the per-event $aicachereporting_exclusive flag, never on provider name — see [cache accounting](./references/cache-accounting.md) for the exclusive-vs-inclusive formula.

distinctid is the canonical user dimension. Customers often attach custom properties (feature, tenantid, workflow_name) — discover them with posthog:read-data-schema before grouping. Don't guess names.

Workflow: total spend in a window

posthog:execute-sql
SELECT round(sum(toFloat(properties.$ai_total_cost_usd)), 4) AS total_cost_usd
FROM events
WHERE event IN ('$ai_generation', '$ai_embedding')
    AND timestamp >= now() - INTERVAL 30 DAY

Workflow: cost breakdowns

Every cost question is a variation of the same template — group by a dimension, aggregate $aitotalcost_usd. See [breakdown patterns](./references/breakdown-patterns.md) for ready-to-run recipes:

• Cost over time (daily)
• Cost by model
• Cost by user (top spenders)
• Cost by trace (top expensive traces)
• Cost by custom dimension
• Cost-per-call distribution
• Input vs output vs cache economics

Workflow: inspect a single trace's cost

When the user pastes a trace URL and asks about its cost, fetch the trace and surface the per-event breakdown:

posthog:query-llm-trace
{ "traceId": "<trace_id>", "dateRange": {"date_from": "-30d"} }

Sum $aitotalcost_usd across the returned events, grouped by span name or model, to show which step(s) drove the cost. The trace response already includes totalCost as a convenience.

Workflow: debug a cost regression

"Our LLM bill jumped — why?" is almost always one of: more calls, bigger prompts, a new model, or a change in cache-hit rate. Work through them in order — see [regression debugging](./references/regression-debugging.md) for the 5-step playbook.

Workflow: materialize as an insight, dashboard, or alert

After ad-hoc queries answer the question, persist them as insights, bundle into a dashboard, or wire up alerts. See [materializing](./references/materializing.md) for ready-to-run JSON for posthog:insight-create, posthog:dashboard-create, and posthog:alert-create.

Constructing UI links

Never hand-write https://app.posthog.com/... links. That host drops the region and the project prefix, so the user is redirected to login instead of the page you meant.

• Prefer the canonical URL the tool returns. query-llm-traces-list and query-llm-trace

return posthogUrl — surface that value. For a single trace, append ?timestamp=<urlencoded_iso> (the trace's earliest event time) to that URL; the returned link carries no timestamp, and without one the trace page scans from a fixed early date instead of the ten-minute window around the trace.

• Otherwise build the link with generate-app-url. It resolves the correct region host and

/project/<id>/ prefix (e.g. https://us.posthog.com/project/2/ai-observability/traces). Pass concrete ids via params, never inline them into the path. - Dashboard: generate-app-url {url: "/ai-observability/dashboard"} - Traces list (sort by cost): generate-app-url {url: "/ai-observability/traces"} - Generations list: generate-app-url {url: "/ai-observability/generations"} - Users list (per-user cost): generate-app-url {url: "/ai-observability/users"} - Single trace: generate-app-url {url: "/ai-observability/traces/{id}", params: {id: "<trace_id>"}}

generate-app-url cannot express query params, so append the ?timestamp=<urlencodediso> described above to a single-trace link yourself.

Always surface a UI link so the user can verify visually.

Keeping this skill current

Provider reporting behavior (which tokens are inclusive vs exclusive, which costs show up where) shifts over time and can differ between SDK versions for the same provider. To avoid rot:

• Branch on event-level flags ($aicachereporting_exclusive,

$aicostmodel_source) rather than hardcoded provider or model names. Those flags are ingestion's resolved answer for the specific event and are the right source of truth.

• $aitotalcost_usd is always authoritative for rollups — prefer it

over summing components, which can drift as new cost categories are added.

• For anything not covered here (new cost categories, changes to

pricing lookup, provider additions), run posthog:docs-search for "calculating costs" or "AI observability" first rather than trusting a hardcoded rule in this file.

• If you find this skill contradicting the UI, trust the UI and flag

the skill for an update.

Tips

• Always set a time range — cost queries without one scan the full events table
• Token, cost, model, and $aitraceid properties are on events — but message content ($aiinput / $aioutputchoices) lives only on the posthog.aievents table; see the traces skill's [event reference](../exploring-llm-traces/references/events-and-properties.md) if you need content alongside cost
• Always include $aiembedding alongside $aigeneration when summing cost; embeddings are cheap per-call but add up at scale
• Costs are written at ingestion (see Calculating LLM costs) — if $aitotalcostusd is missing or zero, read $aicostmodelsource first: passthrough means the SDK supplied costs; custom means custom token prices; openrouter / manual mean automatic lookup; missing means the model wasn't matched (unusual custom model, fine-tune). Grep: countIf(properties.$aitotalcost_usd IS NULL) per (model, source)
• Custom pricing uses per-token prices, not per-million — if a custom-priced model looks ~1M× too expensive or too cheap, that's almost always the bug
• Exclude errored calls from cost totals only when explicitly asked — providers still charge for many error modes, and including them gives the truthful bill
• For per-user totals, exclude rows where distinctid = properties.$aitraceid — some SDKs default distinctid to the trace ID when no user is set
• Cost is additive across $aigeneration + $aiembedding events within a trace; summing on $aispan gives zero. $aitrace may carry $aitotalcostusd from some SDK wrappers — don't include it in rollups or you'll double-count. $aievaluation events also carry cost but are not part of the stock UI rollups; include them only when the user explicitly wants evaluation spend in the total
• Cache-hit rate depends on $aicachereporting_exclusive — branch on the event-level flag rather than on provider or model name. Provider behavior and SDK versions drift; the flag is ingestion's resolved answer for that specific event
• When answering "why is X expensive?", show the cost and the token split — the user almost always wants to know whether to shrink prompts, shrink outputs, or switch models
• Before building a custom dashboard, check whether the stock /ai-observability/dashboard tiles already answer the question — re-creating them is churn
• For large tenants, materialize common cost queries as insights and reuse via insight-query; ad-hoc SQL is fine for one-offs but re-running it on every dashboard load is expensive

References

• [cost properties](./references/cost-properties.md) — full property schema, total-cost rationale, event-set rules
• [cost sources](./references/cost-sources.md) — how costs get set at ingestion plus a diagnostic query
• [cache accounting](./references/cache-accounting.md) — exclusive vs inclusive providers, cache-hit-rate formula
• [breakdown patterns](./references/breakdown-patterns.md) — SQL recipes for every common breakdown
• [regression debugging](./references/regression-debugging.md) — 5-step playbook for cost spikes
• [materializing](./references/materializing.md) — insight, dashboard, and alert JSON

Related skills

• analyzing-expensive-users — who drives the spend, and whether their usage pattern explains it
• exploring-llm-traces — inspect the expensive traces the breakdowns point at