Claude Sonnet 4.6 and Opus quota stuck at "0 to 0" cannot increase

Google Cloud Build with AI Generative AI & Foundational Models
,
1

Hi everyone,

I’m having an issue with Anthropic model quotas on Vertex AI.

My setup:

Billing is active (upgraded from free trial)

Vertex AI API is enabled

Anthropic Marketplace agreement is accepted
The problem:

When I try to increase the quota for anthropic-claude-sonnet-4-6, the allowed range shows “0 to 0”. The error says: “According to your service usage history, you do not meet the requirements for a quota increase at this time.”

Note: My billing account was recently upgraded from free trial, which may be related.
Is this a known restriction? Is there a manual override or escalation path available?
Thank you!

4

smae with mine, stuck at zero, both aws and google, and the annoying thing, you have to pay for premium support before you can be answered

5

Hi,

I have been trying for 3 days to get a quota increase approved for Claude Opus 4-6 on Vertex AI but keep getting rejected with no explanation.

Project ID: (PII Removed by Staff)
Region: us-east5
Model: anthropic-claude-opus-4-6

Quotas requested:

online_prediction_input_tokens_per_minute_per_base_model → 100,000

online_prediction_output_tokens_per_minute_per_base_model → 20,000
The model is GA, billing is active, Vertex AI API is enabled, and the model is enabled in Model Garden. Current quota is 0 which makes the model completely unusable.
Has anyone else faced this? Is there a minimum account age or spending requirement before quota increases are approved?
Thanks

6

Similar issue cant reach out to them nor get the Quota increase to be approved. Any luck or suggestions

7

Greetings.
I too have been trying to get in touch with one of my sales rep that I conversed with regarding Tier increase on the Gemini api, so I applied for increase directly with the Support and I was recommended that my request has been forwarded to the rep and I should also forward the support ticket number for hopefully a faster reply.

But it’s been almost 3-4 weeks since the initial communique and still waiting.
I’m not a Vertex AI user yet but I got Tier 3 on Gemini and have various other deployments and Cloud Run, so I thought this wouldn’t be a long process but Im curious about what’s up.
And I applied for Sonnet 4.6 with the justification that Im already paying for the stream from Github and just wanted to consolidate all AI cloud use into one umbrella, and Sonnet 4.6 is a perfect model for a local production executive Agent.

Patience is virtue I guess. :slight_smile:

8

I have the same issue with Claude Opus 4.7.

Project ID: (PII Removed by Staff)
Model: anthropic-claude-opus-4-7
Region: Global endpoint
Billing: Active paid account (upgraded from free trial on April 24, 2026)
Vertex AI API: Enabled
Model Garden: Claude Opus 4.7 enabled

All quotas show 0 and console says “value must be between 0 and 0”.
Quota increase request is automatically rejected.

Could someone from Google please manually review and enable
quota for this project? I only need minimal quota to start:

  • Global requests per minute: 60
  • Global input tokens per minute: 500,000
  • Global output tokens per minute: 100,000

Thank you.

9

Hi, this behavior is expected, but the error message lacks context.

What you’re seeing with the “0 to 0” quota range is not a temporary glitch or misconfiguration. It indicates that your project is currently not eligible for that model at all, so the quota system is effectively locked rather than limited.

Why this happens

For Anthropic models (e.g. anthropic-claude-sonnet-4-6) on Google Cloud Vertex AI, access is not governed by quotas alone. There is an eligibility layer evaluated before quotas are even considered.

That eligibility is based on a combination of:

  • Billing account maturity (not just “active,” but established)

  • Verified spend history (beyond free trial usage)

  • Prior usage of Vertex AI services

  • Internal trust/risk signals applied to third-party models like those from Anthropic

If those signals are insufficient, the system sets:

  • Minimum quota = 0

  • Maximum quota = 0

Which is exactly what you’re seeing.

Why upgrading from free trial doesn’t immediately fix it

Moving from trial → paid enables billing, but does not instantly move your account into a higher trust tier.

There is typically a delay while the system evaluates:

  • Payment profile stability

  • Initial billable activity

  • Service usage patterns

During this period, partner models can remain inaccessible even though everything appears correctly configured.

What actually resolves this

There isn’t a single “toggle,” but there are three reliable paths:

1) Build usage history inside Vertex AI (most consistent)

Use available Google-native models first (e.g. Gemini) within Google Cloud Vertex AI:

  • Make regular API calls

  • Generate some billable usage

  • Let this run for a few days

This directly addresses the “service usage history” requirement in the error.

2) Submit a support request (not a quota increase)

Quota requests will continue to fail automatically. Instead, go through Google Cloud Support and open a ticket under:

  • Vertex AI → Generative AI → Model access

Include:

  • Project ID

  • Billing account ID

  • Model name (anthropic-claude-sonnet-4-6)

  • Short description of your use case

  • Expected usage pattern

This routes to a manual review path, which the quota UI does not.

3) Use account-level escalation (if available)

If you have access to a Google Cloud account manager or partner channel, they can:

  • Request allowlisting

  • Accelerate trust tier evaluation

This is often the fastest path in production environments.

Additional note on architecture

Anthropic models on Vertex AI are dual-gated:

  1. Google Cloud eligibility

  2. Partner (Anthropic) access controls

So even with:

  • API enabled

  • Marketplace agreement accepted

  • Billing active

Access can still be restricted until both layers are satisfied.

Practical takeaway

The error message is accurate but incomplete. It’s not saying “try again later,” it’s saying:

“This project has not yet met the internal criteria required to access this model.”

The resolution is therefore not to retry the quota request, but to:

  • Establish usage history

  • Or trigger a manual review

10

Hi, having the exact same issue.

Project: (PII Removed by Staff) Region: US multi-region Billing: Active paid account with real spend ($1.09)

All Claude Opus 4.7 quotas stuck at 0. Every quota increase request gets auto-rejected with “does not meet requirements.” Contacted Google Sales twice — no response. This is blocking my production application from launching.

Can a Google team member please help get this unblocked? Happy to provide any additional info.

Thank you

11

I had the same issue, from my understanding Opus uses too much compute so unless Google sees your product as truly worth it, you cant get it active. I’m assuming you working through the steps via that Google page, just use Gemma 4 instead of Opus.

If you think your product is worth the compute, get in touch with sales.

12

Yep, I’ve spent $1200 in past 5 months on gemini tokens and yet I can’t get anthropic to work at all - - -same issue - with quotea -

13

Not a significant enough payment history I sure averaging $300 a month of my personal funds for the pest 4 months is enough but then again I guess they’re used to enterprise customers