This is primarily a bugfix release. The main new features are clipping
support for canvas figures, reporting of configuration entries that have no
effect (useful e.g. when porting models to a new version of a model library),
and the ability of the NED documentation generator to link to the separately
generated documentation of referenced projects and to produce PNG images. The
main bugfix areas are the correctness of recorded and analyzed results (units
of count scalars, field scalars and dB; warm-up period handling), WATCH() and
inspector display, connection drawing in Qtenv, rendering in the graphical NED
editor, and matplotlib and Pandas compatibility in the Analysis Tool and the
Python library.
Issues fixed in version 6.5.0.
Read the What’s New page for more details.
DetailsEver wished you could just ask your simulation why a packet was dropped, or let an AI co-pilot diagnose a network anomaly for you?
We are excited to share the OMNeT++ 6.4.0 AI & LLM Technology Preview 2, bringing the power of Large Language Models directly into your simulation workflow!
We need your help to test these experimental features and shape the future of AI-assisted simulation. Here is what you can do with this new preview:
What’s Inside?
- Chat directly with Qtenv: A new integrated chat interface in Qtenv lets you interact with LLM services (both cloud-based and local models like Ollama) without leaving the simulation window.
- AI-Controlled Simulations: Built on the standard Model Context Protocol (MCP), the simulation hosts an MCP server. This allows LLMs to query and programmatically manipulate your active simulations in real time.
- On-the-Fly C++ Execution (Linux-only): The LLM can generate and run custom C++ snippets using a JIT interpreter to inspect complex data structures and answer highly specific queries. (Launched safely via the
opp_sandboxutility).
See It in Action
We have prepared two practical, hands-on showcases to demonstrate these features:
- Finding a 3-Way Collision in Aloha – Watch an AI agent pinpoint a rare collision event.
- Diagnosing Clock Sync Anomalies in TSN – Witness an AI co-pilot perform root-cause analysis on a Time-Sensitive Network.
How to Try It & Give Feedback
This preview is available for Linux (x86_64) and macOS (aarch64).
- Head over to the Preview Downloads page to download the pre-built packages.
- Read the Release Notes for detailed setup instructions.
- Share your thoughts, bugs, and feature ideas with us on GitHub Discussions!
Your feedback is welcome! Please download it, try the showcases, run it against your own models, and let us know what you think!
DetailsThis release brings major improvements to the simulation kernel’s numerical stability, a redesigned WATCH mechanism with full STL container inspection, a pluggable RNG manager, and significant NED editor enhancements including figure rendering, rename refactoring, and new display string features. The Qtenv inspector and the IDE also received substantial usability improvements.
Issues fixed in version 6.4.0.
Read the What’s New page for more details.
DetailsCheck out the TicToc tutorial, browse the documentation and the model catalog or chat with an AI about OMNeT++. Watch introductory videos, and explore the featured network simulation projects (see banner).
The OMNeT++ Summit website is a great resource if you want to explore what can be done with the simulator. OMNeT++ Summits are yearly online/offline events where community members present their work.
Past events: 2022, 2021, 2020, 2019, 2018, 2017, 2016, 2015, 2014, 2013, 2012, 2011, 2010, 2009, 2008.