Worked examples
The simulator's four example sessions on Gemini 2.5 Flash-Lite with implicit caching, priced turn by turn by the same cost model. Best case: Google publishes no lifetime for implicit caching and does not guarantee hits, so this assumes every eligible turn hits.
-
Coding agent
Tool calls and edits in quick succession, with a few longer pauses. 40 turns, 20K tokens of starting context, 3K in and 600 out per turn, 1.5 minutes apart, three long breaks of 8 minutes.
Open this in the simulator- With caching
- $0.062
- Without
- $0.38
- Saved
- $0.32 84%
- Pays for itself on turn 2
-
Long chat
A person reading and typing, with one long break. 30 turns, 3K tokens of starting context, 150 in and 500 out per turn, 3 minutes apart, one long break of 45 minutes.
Open this in the simulator- With caching
- $0.012
- Without
- $0.044
- Saved
- $0.032 73%
- Pays for itself on turn 2
-
Big document
One large document, questioned a dozen times. 12 turns, 150K tokens of starting context, 100 in and 700 out per turn, 7 minutes apart.
Open this in the simulator- With caching
- $0.036
- Without
- $0.19
- Saved
- $0.15 81%
- Pays for itself on turn 2
-
Unattended loop
A script calling the model nonstop, the prompt growing every turn. 200 turns, 30K tokens of starting context, 1.5K in and 300 out per turn, 30 seconds apart.
Open this in the simulator- With caching
- $0.48
- Without
- $4.24
- Saved
- $3.76 89%
- Pays for itself on turn 2
The coding agent on each cache setting
The same 40-turn session, changing only the cache setting. Without caching it costs $0.38.
| Cache setting | With caching | Saved | Share saved | Pays off on turn |
|---|---|---|---|---|
| Implicit (automatic), best case | $0.062 | $0.32 | 84% | 2 |
| Explicit, 1 hour TTL | $0.35 | $0.028 | 7.3% | 13 |
| Explicit, 5 minute TTL | $0.35 | $0.035 | 9.2% | 3 |
| Explicit, 24 hour TTL | $0.79 | -$0.41 | -107% | Never |
When caching pays for itself
For a prefix sent again and again inside the cache lifetime. The count does not depend on the prefix size, as long as it is above the minimum.
Implicit (automatic)
There is no write charge. A read costs $0.01 per million tokens instead of $0.10, so caching never costs more than sending the prefix uncached.
A cached prefix pays for itself after 1 read.
Best case Google publishes no lifetime and no guarantee of a hit, so this shows the best case: every eligible turn hits.
Explicit, 1 hour TTL
Creating the cache bills the prefix at the $0.10 input price, storage adds $1.00 per million tokens per hour for the whole 1 hour TTL, and each use bills it at $0.01 instead of $0.10.
An explicit cache pays for itself once it is used 13 times within its TTL, counting the request that creates it.
Lifetime Caches only your starting context. Storage is paid for the whole TTL.
Explicit, 5 minute TTL
Creating the cache bills the prefix at the $0.10 input price, storage adds $1.00 per million tokens per hour for the whole 5 minutes TTL, and each use bills it at $0.01 instead of $0.10.
An explicit cache pays for itself once it is used 3 times within its TTL, counting the request that creates it.
Lifetime Caches only your starting context. Storage is paid for the whole TTL.
Explicit, 24 hour TTL
Creating the cache bills the prefix at the $0.10 input price, storage adds $1.00 per million tokens per hour for the whole 24 hours TTL, and each use bills it at $0.01 instead of $0.10.
An explicit cache pays for itself once it is used 268 times within its TTL, counting the request that creates it.
Lifetime Caches only your starting context. Storage is paid for the whole TTL.
Caching rules for Gemini 2.5 Flash-Lite
- Implicit caching is on by default for Gemini 2.5 and newer, with no write charge, no storage charge and no cost-saving guarantee. A hit costs 0.1x the input price. Google publishes no lifetime, so implicit figures here are a best case in which every eligible turn hits.
- Google's implicit caching table does not list Flash-Lite models, so no minimum is published and none is applied.
- Explicit caching lets you create a cache with a TTL, 1 hour by default. Storage costs $1.00 per million tokens per hour for the whole TTL. Creating the cache is billed at the standard input price: Google Cloud's caching overview says so, the Gemini API page does not state it.
- Assumed: explicit caching covers this Flash-Lite model. The docs say "Gemini 3 and 2.5 models".
Sources: Google's pricing page, caching guide, and Google Cloud's caching overview. See all prices and sources.
Similar models
Per million tokens, with the 40-turn coding agent on each model's default cache setting.
| Model | Input | Cache read | Output | Coding agent |
|---|---|---|---|---|
| Gemini 2.5 Flash-Lite | $0.10 | $0.01 | $0.40 | $0.062 best case* |
| Gemini 3.1 Flash-Lite | $0.25 | $0.025 | $1.50 | $0.17 best case* |
| Gemini 3.5 Flash-Lite | $0.30 | $0.03 | $2.50 | $0.22 best case* |
| Claude Haiku 5.5 | $0.10 | $0.01 | $0.50 | $0.37 |
| GPT-6 Luna | $0.10 | $0.01 | $0.50 | $0.068 |
* Gemini implicit caching, best case.
Questions
How much does context caching cost on Gemini 2.5 Flash-Lite?
Input costs $0.10 per million tokens, with no charge to write the cache, $0.01 to read it and $0.40 for output. Explicit caches also cost $1.00 per million tokens per hour of storage.
Is prompt caching worth it for Gemini 2.5 Flash-Lite?
In the 40-turn coding agent example, caching costs $0.062 against $0.38 without, a saving of $0.32 (84%). With implicit caching there is no write charge, so caching never costs more than not caching. This is a best case, because Google does not guarantee hits.
How long does the Gemini 2.5 Flash-Lite cache last?
Implicit (automatic): Google publishes no lifetime and no guarantee of a hit, so this shows the best case: every eligible turn hits. Explicit caching lasts for the TTL you set; the simulator offers 5 minutes, 1 hour or 24 hours. Caches only your starting context. Storage is paid for the whole TTL.
Does Gemini 2.5 Flash-Lite have a minimum cacheable prompt length?
Google does not publish one for this model, so the simulator applies none.
An estimate from published list prices, and your bill may differ. Token counts are what the example sets, not tokenized text. Not covered: batch and fast-mode prices, thinking tokens beyond the output shown, tool fees and regional uplifts.