The Best LLM Observability for OpenAI Apps in 2026, by Use Case

Tracing an OpenAI app is the easy case. Everything here works with the OpenAI SDK, so the integration is rarely the deciding factor. What actually decides it is the stuff underneath: what it costs at real volume, whether you can self-host, and - for one tool - whether the company behind it is still shipping.

That last point matters more than usual in mid-2026, because the tidiest OpenAI-first tool in this set just got frozen. So I judged these four on OpenAI SDK integration, cost at scale, self-host reality, and current vendor status. One of them I am including only to tell you not to start there.

The short version

Tool	Best for	OpenAI fit	Self-host	Starting price
Langfuse	The open-source default	OpenAI SDK integration	Free, MIT	Free / $29/mo
Opik	Cheapest managed cloud	OpenAI integration	Free, Apache-2.0	Free / $19/mo
Portkey	Multi-provider gateway + cost control	Drop-in OpenAI proxy	Gateway only	Free / $49/mo
Helicone	Nobody new - maintenance mode	Proxy, any provider	Free, Apache-2.0, frozen	Free / $79/mo

1. Langfuse - the open-source default

Langfuse is the tool most teams land on, and for an OpenAI app it is a clean fit - it ships an OpenAI SDK integration alongside LangChain, LlamaIndex and OpenTelemetry support, and it logs traces, cost and token usage per call. The reason it is the default is the same reason it wins elsewhere: the self-host actually gets you the real product. Only three features are enterprise-gated, so tracing, evals, prompt management, human annotation and RBAC are all free under MIT.

The economics are the clincher. At 1M events a month Langfuse runs about $101/mo managed, and self-hosting removes the per-trace cost entirely. The catch is operational - Langfuse v3 needs Postgres plus ClickHouse, Redis and S3-compatible storage, four services, and the migration to that architecture is where self-hosters get stuck. It is also a ClickHouse subsidiary after the January 2026 acquisition, MIT license preserved. If you can run the stack, nothing else gives you this much for free.

2. Opik - the cheapest managed cloud

If you want managed hosting and the lowest bill, Opik is the pick. It is Comet’s open-source observability and eval platform, it integrates with the OpenAI SDK among its 60+ framework integrations, and its Pro cloud tier is $19/mo for 100k spans - the cheapest paid cloud of the major eval platforms, with $5 per additional 100k spans. The OSS build is Apache-2.0 with the full feature set self-hosted and no gates, the most permissive license here. There is also a free cloud tier at 25k spans a month, up to 10 members.

The honest gotcha is per-seat pricing at scale - the recurring complaint is that seat costs climb as the team grows, so model it if you are large. A few users report UI slowdown on very large projects. Neither undercuts the core value. For a small OpenAI shop that wants a managed dashboard without running four services, Opik is the best-value entry in the category.

3. Portkey - when the problem is really cost and routing

Portkey is a different animal. It is an LLM gateway that routes to 1,600+ models with fallbacks, caching, budgets and 50+ guardrails, and it happens to include observability. For OpenAI it is a drop-in proxy - point your existing OpenAI SDK base URL at Portkey and you are done, no re-instrumentation - and it auto-enriches OpenTelemetry traces with cost and token metrics per the GenAI conventions. If your problem is “OpenAI costs are exploding and we need budgets, caching and a fallback provider in one place,” this is aimed at you.

But know the split before you plan a deployment. The open-source gateway is Apache-2.0 and self-hosts free - but that gets you routing and a basic dashboard, not real observability. Logs, traces, analytics and retention live on the managed Production tier at $49/mo for 100k logs. The meter caps logs, not requests, so past the cap your traffic keeps flowing but your observability quietly goes dark. Plenty of teams run Portkey as the gateway and feed a dedicated tracer like Langfuse over OTel - that pairing is the right mental model.

4. Helicone - do not start here

I am including Helicone only to tell you not to build on it. It was a clean, open-source, proxy-based observability tool - you changed your OpenAI base URL to route through it and it logged everything, zero SDK. Then Mintlify acquired it on 3 March 2026 and put it in maintenance mode: security patches, bug fixes and new-model support only, no roadmap, and Mintlify is actively helping customers migrate off.

For a new buyer that single fact overrides everything else. The Apache-2.0 self-host still works, but self-hosting a frozen codebase whose own creator is pointing people to the exit is a weak bet. The proxy model was also a structural single point of failure - if Helicone is down, your OpenAI calls fail even when OpenAI is healthy, and every proxied call adds latency (Helicone cites ~10ms). If you are already on it, use the runway to plan your migration. If you are evaluating fresh, skip to Langfuse or Opik.

So which one?

• You want the best open-source platform and can run the self-host stack - Langfuse, roughly $101/mo at 1M events or free self-hosted.
• You want managed hosting for the least money - Opik at $19/mo, or its free cloud tier to start.
• Your real pain is multi-provider cost, routing and budgets - Portkey, and expect to pay for the real observability tier.
• You are on Helicone today - plan your exit. Do not start there fresh.

For the ranked view across the whole category, see our best LLM observability tools roundup, and if you are weighing the gateway question specifically, Portkey vs Langfuse digs into where the two jobs diverge. Every price and date here was read from each vendor’s own materials on 26 July 2026. This category ships breaking changes monthly, so we re-verify every 30 days.