Mario Peng

[Mario Peng @MarioPeng242578]

@adomaswastaken @densechat We show tokens saved so you can extrapolate to other pricing models, and understand transparently the math beind our savings calculation

[Mario Peng @MarioPeng242578]

@AnExiledDev @densechat Regarding quality benchmark, we have some behind-doors tests, which are not as interesting. An open source, transparent benchmark is coming soon 👀

[Mario Peng @MarioPeng242578]

@AnExiledDev @densechat Yeah, breaking cache is unavoidable for any sort of compaction. The question is when to compact and how much, and how many turns after will you break even and start saving compared to a counter factual perfect cache trajectory. This is baked in our compaction strategy

[condense.chat @densechat]

We're giving you 100M free tokens to prove your agent is wasting context.

Most of what your agent sends upstream is dead weight. We built state-of-the-art compaction to strip it, and today we published the receipts.

One real session, run to full depth, benchmarked against headroom:
condense: 53.8% of tokens saved, 37.3% off the bill
headroom: 15.1% and 13.8%

Everything is open in the post. One curl line, and triple your Fable usage.

Full breakdown in the comments