LangSmith Pricing Explained (2026) - Why the Bill Explodes at Scale

LangSmith is the most turnkey observability you can point at a LangChain app. The tracing is genuinely the best available for that stack. The pricing is also the reason “LangSmith alternatives” is a search term. Let me decode exactly why the bill explodes, then show you the number at scale.

The pricing model, decoded

LangSmith has three tiers, and the trap is that two meters stack on top of each other.

Tier	Price	Included traces	Overage	Self-host
Developer	$0	5k base traces/mo, 1 seat	pay-as-you-go	No
Plus	$39/seat/mo	10k base traces/mo	pay-as-you-go	No
Enterprise	Custom	Custom	Custom	Yes

You pay per seat, and then you pay per trace on top. The quota unit is the “base trace,” and base traces cost $2.50 per 1,000 at 14-day retention. Extended traces, which keep 400-day retention, cost $5.00 per 1,000. Upgrading a base trace to extended adds another $2.50 per 1,000.

So there are really three ways the meter climbs - more seats at $39 each, more traces at $2.50 per 1,000, and longer retention that can double your per-trace cost. Deployment for LangGraph agents adds its own compute meters on top (LCU at $1.50, LSU at $1.00).

The Developer free tier is one seat and 5k traces - fine for a prototype, useless for production. Plus at $39 a seat with 10k traces is where most teams start, and where the trace overage quietly takes over the bill.

Estimate your bill

Here is the number that drives migrations. On the Plus plan, one seat, 14-day base retention.

• 1M base traces a month. That is 990k traces of overage past the included 10k, at $2.50 per 1,000, so about $2,475 in trace charges plus the $39 seat - roughly $2,514/mo. This is LangSmith’s real cost at production volume on a single seat.
• Add a second retention need. If those traces need 400-day retention for audit, you are at $5.00 per 1,000, which roughly doubles the trace line toward $5,000/mo.
• Add seats. Every teammate is another $39, and a 10-person team adds $390 before a single extra trace.

The killer is that the trace charge is per-request, uncapped, and tied to how busy your app is. A successful product traces more, so the tool that helps you scale costs more exactly as you scale. And you cannot self-host your way out, because self-hosting is Enterprise-only.

The trap - no self-host escape hatch

This is what makes LangSmith different from the open-source options. LangSmith is fully closed source. There is no open-source build, no free self-host, no MIT-licensed core. Self-hosted and hybrid deployment exist only on Enterprise, behind a contact-sales conversation.

Why does that matter so much? Because self-hosting is the standard escape hatch from trace-volume pricing everywhere else in this category. With an open-source tool, when the managed bill gets ugly, you run it yourself and the per-trace cost disappears. With LangSmith you are stuck on the meter or stuck negotiating an Enterprise contract. When LangChain cut its old generous free tier, the fallout was visible - Reddit threads pushed a wave of users toward self-hosted Langfuse for exactly this reason. Full detail is on the LangSmith page.

Cheaper picks, and when they win

Two alternatives cost a fraction, and the honest case for each is different.

Langfuse is the direct swap and the cheapest at scale. It is framework-agnostic, MIT-licensed, and runs about $101/mo at 1M events managed - the roughly 25x gap that decides most LangSmith migrations. Self-host and the per-trace cost disappears entirely. The catch is operational, not commercial - the self-host is now four services and the v3 migration bites. If you want the same job for a fraction of the money and can run the stack, this is the move. The one thing you give up is LangSmith’s zero-config depth on LangChain and LangGraph code specifically.

Opik is the cheapest managed cloud if you do not want to run anything. Comet’s platform is $19/mo for 100k spans, with $5 per additional 100k, and its OSS build is Apache-2.0 with the full feature set self-hosted. If the appeal of Langfuse is the price but you do not want a four-service self-host, Opik is the lowest-ops way to escape the LangSmith meter. Watch the per-seat cost as your team grows.

Bottom line

LangSmith pricing is cheap to start and brutal at scale, and the two things that make it brutal are the per-trace meter stacking on per-seat charges and the fact that you cannot self-host below Enterprise to escape either one. If you live in LangChain and the bill does not scare you, the integration is worth it. If price is why you are here, Langfuse at roughly 25x cheaper or Opik at $19/mo will serve you for a fraction.

All figures here were read from each vendor’s own materials. This category ships breaking changes monthly, so we re-verify every 30 days. The 25x gap between LangSmith and Langfuse at 1M traces has held for a while - if cost is the question, it is not close.