Reddit: 185:1 token ratio on long agentic tasks despite harness optimization
flq06 · reddit · 2026-09-20
A Reddit user reports that despite heavy harness optimization, long agentic tasks still yield input/output token ratios as high as 185:1, with persistent prefill latency issues. They ask what ratios others see, highlighting a common pain point in long-horizon agent workflows: context bloat and latency.
More from coding & agent
- Jev's instant compression scores each tool call to trim agent context without summarization — FinanceYF5 · 2026-09-20
- Stop using LLMs for ticket triage: Jev turns unstructured input into an executable smart if — sven_ai · 2026-09-20
- Jeff Dean's 1-Hour AI Engineering Lecture: From LLM Basics to Agent Graphs — irinarish · 2026-09-20
- Multi-agent systems work best with clear roles, not more agents — _jaydeepkarale · 2026-09-20
- Jev founder: all AI models are built for human-in-the-loop, not true automation — hardimanjames · 2026-09-20
- Karpathy's Stanford AI engineering lecture: from LLM to Graph in five steps — irinarish · 2026-09-20