Hermes Agent prompt caching now offers four settings, and this comparison shows which one to pick for your workload after the 20 November 2026 v0.23.3 update.

Get the Backlink Gameplan →

The update put one controller in charge of warming, compacting and expiring the cache, and made auto the default, per the release notes.

Below: the settings compared, how providers differ, common mistakes and who each setting suits.

Four days earlier, v0.23.2 had already moved every session onto a single cache clock.

cache_ttl options compared

SettingWhat it doesBest for
"auto"5m for typed chats until the first pause over 5 minutes, then 1h; machine jobs stay 5mAlmost everyone
"5m"Always the short tierFast back-and-forth sessions and batch jobs
"1h"Always the long tierSessions with long gaps every time
"off"No cache markersOAuth plans billing cache writes, proxies with their own cache_control

Source: Hermes configuration docs.

Why auto wins for most users

On Claude a 1-hour write costs 2x base input and a 5-minute write 1.25x (Anthropic docs).

Auto pays the premium only after you have actually paused, so a chat that never pauses never pays for the long tier.

The Hermes docs report roughly 40 percent lower interactive cache-write spend on one install using auto.

Providers compared

Provider routeBehaviour
Claude via Anthropic, OpenRouter, Nous PortalMarked caching at 5m or 1h
Nous PortalFull 1h lifecycle since v0.23.3
Qwen Cloud5-minute cap
OpenAI-compatibleAutomatic, treated as 30 minutes
OpenAI Responses, openai-codex, xAINever warmed
Local modelsNo clock expiry

🔥 Compare model and caching setups with other operators Inside the AI Profit Boardroom you get step-by-step video trainings, the installable Agent OS, four coaching calls a week and 3,400+ members. → Join the AI Profit Boardroom

Warming and compaction at a glance

  • Warm at 5/6 of the lifetime with a one-token request.
  • Up to 3 idle keep-alives per message you send.
  • Unlimited warming during a running tool call, capped by cost.
  • Compact idle chats with a window of 30 minutes or more when the context is at least half the threshold.
  • Offload old tool results to disk when the cache expires.

Common mistakes

  1. Pinning "1h" for subagents; they are clamped to 5 minutes anyway.
  2. Leaving an old "5m" pin in config; the v54 migration only removes the old template line, so check yours.
  3. Expecting warming on Responses API routes, which are never warmed.
  4. Ignoring /context, which shows whether the tier is proven.
  5. Using budget-based extended thinking and expecting keep-alives; those requests are never warmed.

Rank in AI Search or It's FREE →

Who each setting suits

  • Auto: solo operators and teams mixing chat and automation.
  • 5m: batch pipelines and rapid chats.
  • 1h: analysts who return to the same long session every half hour.
  • Off: specific billing or proxy setups.

Pair the setting with a model from our best Hermes Agent model roundup.

What changed in v0.23.3, at a glance

ChangeBeforeAfter
Long tool runsCache could expire mid-runKept warm until the tool returns
Idle chatsExpired cold, full re-send on returnCompacted before the one-hour cache expires
Nous PortalShorter lifecycleFull one-hour lifecycle
cache_ttl defaultFixed tierauto

Source: the v0.23.3 release notes, which describe one prompt-cache lifecycle controller.

What breaks the cache, compared

ActionCache effect
Adding a new messagePrefix kept; rolling window moves forward
Editing the system prompt mid-chatWhole prefix rewritten
Inserting or removing earlier messagesCache lost from that point on
/model switch, fallback or credential rotationZero cache hits on the next request
CompressionCompressed region recached; system prompt cache survives

All rows come from the Hermes developer guide on compression and caching.

Session types compared

Session sourcePaceTier under auto
CLI, TUI, desktop, chat platformsHuman-pacedCan use the one-hour tier
Subagents and delegated childrenMachine-paced5 minutes
Cron, webhook, Kanban, API server, batchMachine-paced5 minutes
One-shot commandsMachine-paced5 minutes
Side agents from /background, /btw, /refineSide agent5 minutes

The one-hour tier costs more to write, so it only pays when turns are spaced out, which is why machine-paced sources stay short.

Cache breakpoint strategy

Anthropic allows four cache_control breakpoints per request.

Hermes places one on the system prompt and three on the last three non-system messages, a strategy the guide names system_and_3.

That keeps the stable part cached for the whole session while the recent turns roll forward.

Compression layers compared

LayerFires atPurpose
Agent context compressor50 percent of context (default)Normal context management inside the tool loop
Gateway session hygiene85 percent of contextSafety net for sessions that grew between turns
Cache-edge compaction (new)When idle keep-alives run outShrink the chat before the cache expires

Join the AI Profit Boardroom →

Who benefits most

  • People running long, multi-turn sessions on Claude models.
  • Operators whose agents run long tool calls, such as large file or research jobs.
  • Anyone who parks chats and returns later in the hour.

Short single questions gain little, because there is little prefix to reuse.

Spec routes compared

Route specHow the TTL is setRefresh on read
Anthropic-spec (cache markers present)5m or 1h as requested, capped by route ceilingPer provider rules
OpenAI-spec (all other cached requests)Declared 30 minutesExtends to read time plus 30 minutes
OpenAI-spec with 24h retentionWarming deadline 30m, estimated end 24hNot pruned cold on the clock
Local endpointsNo timed expiryNot applicable

All of these are described in the developer guide.

Cache confidence levels compared

Level in /contextMeaningHermes behaviour
ProvenThe response named the tierWarms and compacts at the cache edge
DeclaredWhat Hermes requestedActs on the requested tier
Assumed or unknownNo confirmationWill not warm or compact at the edge

"Unknown" often follows a model switch, fallback or compaction that moved the cache.

How to verify savings

Anthropic's API reports cache_creation_input_tokens and cache_read_input_tokens on every response, per its prompt caching docs.

On a warm session most input lands in cache reads.

A turn heavy on cache creation means the prefix was rewritten, usually by a model switch or a prompt edit.

Compare a week before and after the update on the same workload for a fair read.

Pinning vs auto: decision table

Your patternBest setting
Mixed interactive and automated useauto
Every session has gaps over five minutes"1h"
Rapid machine-paced jobs only"5m" or auto
OAuth plan billing cache writes as extra usageoff
Proxy that adds its own cache_controloff

Options and their behaviour are listed in the configuration docs.

Warming rules compared

StateWarming behaviour
Idle chatAt most three keep-alives per message you send
Tool runningKeeps warming until the tool returns, bounded by cost
Two misses in a rowWarming switches off for the session
Budget-based extended thinkingNo keep-alives

Get the Backlink Gameplan →

Common mistakes

  • Switching models mid-session to compare them, which zeroes the cache each time.
  • Editing the system prompt or skills during a long chat.
  • Pinning one hour for cron or subagent work.
  • Reading "unknown" in /context as a bug when it follows a fallback.
  • Leaving caching on behind a proxy that adds its own cache markers.

Before and after scenarios

ScenarioBefore v0.23.3After v0.23.3
A tool runs for longer than the cache lifetimeCache could expire; next turn re-sent in fullKept warm until the tool returns
You leave a chat for 40 minutesCould expire coldIdle keep-alives, then compaction before expiry
You return after the cache expiredFull history re-sentOld tool results already offloaded, chat smaller
Nous Portal sessionShorter lifecycleFull one-hour lifecycle

The pattern is the same in each row: fewer full-price re-sends.

Glossary

TermMeaning
TTLHow long the provider keeps the cached prefix
BreakpointA cache_control marker telling the provider where to cache
Keep-aliveA tiny replay that refreshes the cache before it expires
Cache edgeThe moment the cache is about to expire
Route ceilingA cap on how long Hermes assumes a provider keeps the cache
OffloadMoving old tool results to disk with a pointer left in place

Quick picks

  • Most users: update and leave cache_ttl on auto.
  • Long gaps every session: consider pinning "1h".
  • Cache writes billed as extra usage: turn caching off.
  • Provider keeps less than asked: set a route_ttl_ceilings entry.
  • Local models: nothing to do; no timed expiry.

Options are documented in the configuration docs.

Checklist after updating

  • Hermes on v0.23.3 or newer.
  • cache_ttl on auto or unset.
  • /context shows a warm cache on a Claude session.
  • No mid-session model switches in your usual workflow.
  • Usage logs show most input as cache reads on long sessions.

Session habits compared

HabitCache outcome
One model for the whole sessionPrefix reused every turn
Skills loaded at the startSystem prompt stays stable
New session per topic with /newSmall, relevant prefix
Switching models to compare answersZero cache hits after each switch
Editing skills mid-chatPrefix rewritten from the change
One giant chat for unrelated jobsLarge prefix, more compaction, more rewrites

Rank in AI Search or It's FREE →

The top three habits are free and matter more than any config change.

Compression and the cache

When Hermes compresses a long chat, the compressed region loses its cache but the system prompt cache survives.

The rolling window of recent messages rebuilds caching within one or two turns, per the developer guide.

So compression costs a little once, then savings resume quickly.

That is why cache-edge compaction is a net win: it trades one small rewrite for avoiding a full cold re-send later.

Who should change nothing

If you use Claude through Anthropic or OpenRouter, run mixed chats and background jobs, and have never touched prompt_caching, you are already on the best setting after the update.

The only action is to update Hermes and keep the session habits above.

Where to look when costs rise

  1. Run /context and check whether the cache shows warm and proven.
  2. Check whether a model switch, fallback or credential rotation happened in the session.
  3. Check whether skills or the system prompt changed mid-chat.
  4. Check whether the session is machine-paced and therefore on the five-minute tier.

One of those four explains most cost spikes.

Verdict

Stay on auto, read /context once a week, and learn the wider agent stack in the AI Profit Boardroom.

Join the AI Profit Boardroom →

FAQ: Hermes Agent Prompt Caching Compared

Is auto better than 1h?

For most users yes, because it pays the 1-hour premium only after you pause.

Do subagents use the 1-hour cache?

No, delegated subagents are always clamped to 5 minutes.

Does OpenAI get warmed?

OpenAI-compatible routes are warmed on a 30-minute assumption; Responses API sessions are never warmed.

Where do I check the tier in use?

Run /context in the session.

📺 Video notes + links to the tools 👉

🎥 Learn how I make these videos 👉

🆓 Get a FREE AI Course + Community + 1,000 AI Agents 👉

About Julian

I'm Julian Goldie, an SEO entrepreneur, author and founder of the AI Profit Boardroom, which has 3,400+ members.

I help business owners scale with AI agents, automation and SEO.

  • I've built a 7-figure SEO and link building agency, Goldie Agency, with a team of 70+ SEO experts.
  • I've grown a YouTube channel to 400,000+ subscribers.
  • I wrote "Link Building Mastery".
  • My Udemy courses have taught over 29,000 students.

→ Get my best AI training inside the AI Profit Boardroom

Check your config.yaml for an old pin, set auto if you find one, and compare costs after a week.

For guided setups, join the AI Profit Boardroom, and keep this comparison for every Hermes Agent prompt caching decision.