Has anyone here managed to get context caching working for gemini-2.0-flash-001 on a free tier account?

Gemini API
,
1

Hi there,

According to this pricing page, the context caching is supported for gemini-2.0-flash-001 even for free tier account:

I tried it, but getting the following error:
google.genai.errors.ClientError: 429 RESOURCE_EXHAUSTED. {‘error’: {‘code’: 429, ‘message’: ‘TotalCachedContentStorageTokensPerModelFreeTier limit exceeded for model gemini-2.0-flash: limit=0, requested=116251’, ‘status’: ‘RESOURCE_EXHAUSTED’}}

Please advise. Thank you.

3

Hi @limcheekin, You will face this error when you send too many requests per minute with the free tier Gemini API. Please try after some time. Thank You.

5

image
image1607×276 36.1 KB

@Kiran_Sai_Ramineni Here’s this stat from an account with all 0’s
From the first request ever with this account
i get the same error as above (only when using cache context)

Any suggestions ?

6

Thanks for quick response.

I faced this error from the first request too.

7

Thanks for sharing the screen.

Where could I find this screen?

9

Hi @Ayoub_Kazar @limcheekin, May I know the number of tokens you are trying to cache. Also could you please confirm you are facing this error while creating the cache or while making a request on the cached content. Thank You.

11

I am facing this error when the app creating the cache.

I think the number of tokens is stated in the error message above ‘requested=116251’.

Hope this clarify.

Thank you.

13

Hi, I try to create a cache for gemini-2.0-flash, but I always receive HTTP 503, do you meet the same issue? Thank you!

14

This is on creating the cache
Here’s the number of tokens: 298263 tokens

429 RESOURCE_EXHAUSTED. {'error': {'code': 429, 'message': 'TotalCachedContentStorageTokensPerModelFreeTier limit exceeded for model gemini-2.0-flash: limit=0, requested=298263', 'status': 'RESOURCE_EXHAUSTED'}}
15

No, recheck the docs on how to create a cache, then maybe you’ll reach our issue too lol

16

@Long_Peng @Kiran_Sai_Ramineni Please checkout the following code I used to create the cache at TelegramGPT/gemini.py at main · limcheekin/TelegramGPT · GitHub. The same code working for gemini-1.5-flash-002.

Hope this help to resolve the issue faster. Thanks.

18

Got the same issue.

My account is on free tier.

19

Hi @Long_Peng, @Ayoub_Kazar @limcheekin While reproducing i have also received the same error while using the free tire account. will escalate this issue to the engineering team. Thank You.

20

Great! Please keep us posted once the issue has been resolved.

Thanks.

22

@Kiran_Sai_Ramineni May I know any update for this issue?

24

Hello @Kiran_Sai_Ramineni ,
do we have a workaround/solution for this issue?

I am currently encountering this issue and I have a deadline coming up where I wanted to use the 2.0 Flash model.

Or if anyone managed to get explicit context caching working please let me know.

26

Hi! Any updates on this? I am facing the same issue now

28

Any updates on this?

30

The issue is still here. I had a Similar Issue about two months ago with gemini flash 1. Somehow the documentation is also not always correct/up to date.

32

I have listed it on the issue tracker Google Issue Tracker