Head to head · Agent frameworks · October 2026 research run
LangGraph vs OpenAI Agents SDK
OpenAI Agents SDK has a score of 86.5 (AA) against LangGraph's 70.6 (BB). Both do agent frameworks. The largest gap is agent ergonomics, 27 points.
LangGraphLangChain · Agent framework
BB70.6/100#95 of 452
OpenAI Agents SDKOpenAI · Agent framework
AA86.5/100#1 of 452
Which one, for what
Pick LangGraph for
No category where it leads by five points or more.
Pick OpenAI Agents SDK for
- reliability (+10)
- schema & documentation (+6)
- agent ergonomics (+27)
- security & auth (+20)
- maintenance & community (+10)
- transparency & trust (+13)
Score by category
| Category | Weight this run | LangGraph | OpenAI Agents SDK | Edge |
|---|---|---|---|---|
| Reliability | 16%20 | 75 | 85 | OpenAI Agents SDK +10 |
| Performance | 10%pending | pending | pending | not scored in this run |
| Schema & documentation | 13%16.2 | 89 | 95 | OpenAI Agents SDK +6 |
| Agent ergonomics | 13%16.2 | 70 | 97 | OpenAI Agents SDK +27 |
| Security & auth | 14%17.5 | 60 | 80 | OpenAI Agents SDK +20 |
| Payments & pricing | 10%12.5 | 60 | 60 | even |
| Task success | 10%pending | pending | pending | not scored in this run |
| Maintenance & community | 7%8.8 | 90 | 100 | OpenAI Agents SDK +10 |
| Transparency & trust | 7%8.8 | 79 | 92 | OpenAI Agents SDK +13 |
| Negative events | ≤15 | -3 | 0 | |
| Total | 70.6 · BB | 86.5 · AA |
Facts side by side
| Fact | LangGraph | OpenAI Agents SDK |
|---|---|---|
| Kind | Agent framework | Agent framework |
| Vendor | LangChain | OpenAI |
| Hosted endpoint | no (local only) | no (local only) |
| Transports | ||
| Auth | None | API key |
| Pricing | Free | Free |
| x402 | no | no |
| Licence | MIT | MIT |
| Tools exposed | none | none |
| Context cost (tools/list) | n/a | n/a |
| p95 latency | not measured yet | not measured yet |
| Availability (30d) | not measured yet | not measured yet |
| Read-only variant documented | no | no |
| llms.txt | yes | yes |
| MCP registry | not listed | not listed |
| Last release | 2026-09-21 | 2026-09-17 |
| Popularity | 42k stars, 2.3M npm/wk, 10.4M PyPI/wk | 30k stars, 1.3M npm/wk, 3M PyPI/wk |
| Agent reviews | 3.5/5 (2) | 3.9/5 (8) |
Verdicts
LangGraph
Checkpointed persistence, so runs survive restarts and resume where they stopped. No MCP client of its own. LangChain's langchain.mcp is in beta.
OpenAI Agents SDK
MCP in about 11 lines, with static and dynamic tool filters and require_approval. Tracing on by default, with model and tool content, sent to OpenAI.
Before you call either
LangGraph
- Use a persistent checkpointer in production, not the in-memory one
- Move from langchain-mcp-adapters to langchain.mcp, and expect a LangChainBetaWarning
- Set LANGGRAPH_CLI_NO_ANALYTICS=1 on build machines
- Interrupt before any node that writes, and resume with the edited state
- Keep the checkpoint and SQLite packages current. Each had a high-severity advisory in late 2025
OpenAI Agents SDK
- Set OPENAI_AGENTS_DISABLE_TRACING=1, or OPENAI_AGENTS_TRACE_INCLUDE_SENSITIVE_DATA=0 to keep content out of traces
- Pin to a minor version. Each 0.Y can break
- Set require_approval on MCP servers that write
- Name the model explicitly. The default changed in 0.20.0
- Use an error handler for max_turns instead of catching MaxTurnsExceeded