Hermes, Claude Code, and Codex ran an identical model. Token use
Points: 2 # Comments: 0 Hermes, Claude Code, and Codex ran an identical model. Hermes, Claude Code, and Codex ran an identical model. Token use
By Dillip Chowdary • Sep 02, 2026 • Source: HN Claude/Codex/Fable
What happened
A recent analysis published on The New Stack revealed a massive discrepancy in how different AI agent frameworks consume resources. When running an identical underlying language model, three distinct agent harnesses—Hermes, Claude Code, and Codex—demonstrated a staggering seventy-fold variation in total token usage. This massive spread highlights how the scaffolding, developer prompts, and loop architectures wrapping a model can quietly inflate operating costs without necessarily improving the final output. The discovery has
Always stay in-the-know by getting the most important news and exclusive content delivered fresh to your inbox to learn more about at-scale software development. When you're done, simply close that tab and continue with this form to complete your subscription.
How it works
Please answer a few simple questions to help us deliver the news and resources you are interested in. FIRST NAME REQUIRED LAST NAME REQUIRED COMPANY NAME REQUIRED COUNTRY REQUIRED Select ...
Advertisement
Tech Pulse Daily
Get tomorrow's pulse first
Join engineers who read Tech Pulse before stand-up. Free, weekday mornings.
Why it matters
Wallis and Futuna Western Sahara Yemen Zambia Zimbabwe Åland Islands ZIPCODE REQUIRED Great to meet you! See the full write-up from HN Claude/Codex/Fable via the source link for quotes and complete context.
Who is affected
Read the original coverage at HN Claude/Codex/Fable via the source link above for the complete details and primary quotes.
What to watch next
Cross-check release notes and official docs before changing production systems based on early reporting.
Developer Action Items
- ☐ Diff the official changelog for Claude / Codex before you bump — APIs, defaults, and removed flags only.
- ☐ Install through the vendor's documented channel in staging; keep a one-command rollback and time-box the canary.
- ☐ Grep your repo for old flag names, lockfile pins, and plugin versions that the notes mark as breaking.
- ☐ Prefer the first patch cut over the day-zero tag unless you have a reason to be on the leading edge.
- ☐ If HN Claude/Codex/Fable did not name a region, plan, or SKU, screenshot the official availability line before you promise it to users.
Author
Dillip Chowdary
Writes Tech Bytes coverage of AI, engineering, and the tools that actually ship. Editor of Tech Pulse Daily.
Related on Tech Bytes
Anthropic launches Claude Fable 5.1 and says it’s up to 45 percent cheaper for agentic…
Read →
John Ternus sends first memo as Apple CEO teasing a ‘huge launch next week’
Read →
Were Claude's Weekly Limits Cut in Half? (and obscured by another 50% "deal"?)
Read →
AfterQuery reportedly becomes Y Combinator’s fastest-ever unicorn, now valued at $3.2B
Read →
Today's Tech Pulse briefing
Full briefing →
Advertisement