Coding-agent harnesses cost more without better benchmark scores
Harness swaps changed outcomes about as often as reruns, while system prompts made some tasks cost up to three times more.
6
Stories per day · 30 days
2 RSS
Relatedclaude-code2coding-agents2ai-agents1backdoor1benchmarks1codex1data-poisoning1github-copilot1instruction-tuning1llm-trojan1opencode1swe-bench1tinyfish1web-search1
6 stories, newest first · one card per story
Stories tagged #token-cost
TinyFish post says search is free for AI agents
The fetched page is a cookie notice, so token-cost claims are not available.
3
LLM Trojans that increase response lengths and cost
The authors also describe a defense and a trigger that fires on semantically similar instructions.
TokenEconomizer cuts coding-agent token costs on sequential tasks
Authors report 18.4-30.9% lower average token cost on Claude Code, Codex, and GitHub Copilot CLI with comparable success.
Where Do the Tokens Go? Understanding and Reducing Costs in LLM Agents for Vulnerability Discovery
A 200-trace CyberGym study finds reasoning dominates token cost, and AVRI lowers it.
AI agent swarms are a massive waste of tokens with zero quality gain, says OpenAI Codex developer
OpenAI Codex developer Eric Provencher warns that running more than two parallel AI sub-agents burns tokens without improving output quality.