ReasonBlocks
The runtime layer that makes AI agents cheaper and more reliable
About ReasonBlocks
ReasonBlocks is the runtime layer that makes AI agents cheaper and more reliable. We correct agent reasoning and compress tokens mid-run, with gains compounding across runs. On SWE-Bench pro we have seen a 52% token reduction and 42% accuracy lift.
Public traction evidence
Each signal links to the public source used for attribution.
- X
Paul Graham post about ReasonBlocks cutting LLM token costs by about half
Paul Graham described doing office hours with a startup that cuts companies' LLM token costs by optimizing requests, cutting costs by about half. The official YC Spring 2026 snapshot identifies ReasonBlocks as the runtime layer that makes AI agents cheaper and cites a 52% token...
- LinkedIn
We're launching ReasonBlocks!
We're launching ReasonBlocks! Every AI agent call starts from scratch — your agent solves a problem and immediately forgets how it got to the solution. Next time it sees the same type of task, it explores from zero again. Multiply that across code gen, finance dashboards, pen...
- LinkedIn
ReasonBlocks (YC P26) just got featured in Forbes as one of the most promising YC companies in the batch!
ReasonBlocks (YC P26) just got featured in Forbes as one of the most promising YC companies in the batch! We're the runtime layer that makes AI agents cheaper and more reliable: 42% more accurate and 52% cheaper on SWE-Bench Pro, all from a lightweight SDK. If your agent costs...