Verified pricing
LangSmith pricing
LangSmith uses a hybrid pricing model with monthly per-seat fees and usage-based charges for traces, deployments, and fleet runs.
Developer plan includes 1 seat, 5,000 base traces per month, and 50 Fleet runs per month for $0.
Tiers
- 5k base traces / mo
- Tracing to debug agent execution
- Online and offline evals
- Prompt Hub, Playground, and Canvas
- 1 Fleet agent
- 50 Fleet runs / mo
- 10k base traces / mo
- 1 dev-sized agent deployment runs included
- Unlimited Fleet agents
- 500 Fleet runs / mo
- Up to 3 workspaces
- Email support
- Custom trace volume
- Hybrid and self-hosted options
- Custom SSO and RBAC
- Support SLA
- Team trainings & architectural guidance
- Dedicated engineering team access
Hidden costs
Costs that do not appear in the headline price. Sourced from the same pricing pages and product docs as the tiers above.
- Base traces (14-day retention): $2.50 per 1k traces (after free allotment)
- Extended traces (400-day retention): $5.00 per 1k traces
- Additional Fleet runs: $0.05 per run
- Production deployment uptime: $0.0036 per minute (~$155/mo)
- Development deployment uptime: $0.0007 per minute (~$30/mo)
- Additional deployment runs: $0.005 per run (after 1st free dev deployment)
- Sandbox compute: $0.0576/vCPU-hr, $0.0185/GiB-hr, $0.000123/GiB-hr
- Model costs: Billed separately by LLM providers
Contract terms
Developer and Plus plans are billed monthly. Seats are pro-rated when added mid-month, but no credits are provided for seat removal mid-cycle. Enterprise plans are billed annually via invoice. Usage (traces, deployment runs) is billed monthly in arrears.
Discounts
- LangSmith for Startups: Discounted rates and credits for VC-backed startups.
How LangSmith compares
LangSmith is priced higher than open-source alternatives like Langfuse, utilizing a per-seat model ($39/seat) combined with usage-based fees. It is considered the gold standard for teams already within the LangChain ecosystem, whereas enterprise-grade platforms like Arize or Braintrust may offer different scaling dynamics for non-LangChain workloads.
Alternatives by price
Who this fits
Best for professional teams building complex agentic applications with LangChain or LangGraph who require robust debugging, evaluation, and managed deployment infrastructure.
Notes
- Traces with feedback are automatically upgraded to extended retention (400 days).
- Plus plan allows up to 3 workspaces.
- Enterprise tier supports hybrid and self-hosted deployments in VPC.
- 1st dev-sized deployment includes unlimited runs, but uptime is still charged.
Get a real quote
Prices above are list. Enterprise contracts almost always settle below list with annual commitment, multi-product bundles, or volume tiers. Send LangSmith a quote request through AI Compliance Vendors and the conversation starts with the right context.
Our pricing methodology
Every tier on this page cites a public source. We do not paraphrase analyst reports or print numbers we cannot link to. If a tier exists but its price is hidden, we mark it "Contact vendor" rather than guess. If you spot a stale price, tell us and we will re-verify within 48 hours.
For context across all 10 vendors with public pricing, read AI Compliance Vendor Pricing in 2026.