I would love a higher context window in Lumo. At the moment it caps out at 128k, which is not a lot when you upload certain files, and especially now with Lumo 2.0 Max (GLM 5.2) using so many tokens for reasoning. I would say at least 256k would be okay.

I understand that with a higher context window you would need more compute, but that is why I believe this should be made available only to paying customers in a way that won't be burning cash for Proton.

The reasons this is essential are:

1. Discussions - I have had lumo try to compact my conversation many times after only a few rather big questions without even uploading anything.
2. Files for analysis - I tried letting Lumo analyze logs from a bug report one time, it just refused to because the estimated input tokens used by those files were 250k, which is way above the current 128k cap

Alternatively for file analysis, the LLM should have a more efficient tool for parsing, rather than just dumping the entire text file into the prompt for inference. That could help with the limited context length.

At the moment, with Lumo 2.0, the Max model can support up to 1M input tokens but is capped at 128k, I'd love to be able to take advantage of this. I understand that takes a lot more compute, but for many serious tasks this is essential.

I saw a similar post for this, but it was very outdated so I made a new post.