›_ Claude Code Issues
Claude Code Issues / #64398

[BUG] Pro plan blocked by "Usage credits required for 1M context" — no workaround works, quota available

Resolved 💬 6 comments Opened Jun 1, 2026 by kawoatsb-alt Closed Jun 4, 2026

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been reported yet
  • [x] This is a single bug report (please file separate reports for different bugs)
  • [x] I am using the latest version of Claude Code

What's Wrong?

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been adequately resolved
  • [x] This is a single bug report
  • [x] I am using the latest version of Claude Code

Bug Description

Claude Code on the Pro plan blocks every single prompt with the following error, and none of the documented workarounds resolve it:

API Error: Usage credits required for 1M context · turn on usage credits at claude.ai/settings/usage, or use --model to switch to standard context

The error fires at the authentication level, before any model flag or environment variable is processed, making Claude Code completely unusable on the Pro plan. This is a hard blocker.

This Is Not a Quota Issue

At the moment the error occurred, my Account & Usage panel showed plenty of remaining quota:

| Limit | Used | Remaining |
|-------|------|-----------|
| Session (5hr) | 17% | 83% |
| Weekly (7 day) | 15% | 85% |
| Weekly Sonnet | 14% | 86% |

Plan: Claude Pro

Every limit is far from exhausted, yet Claude Code is fully blocked. This is a feature/billing gate firing incorrectly, not a genuine usage limit.

This Worked Before

This is a regression. It began after a recent Claude Code update silently changed the default model to one using a 1M-context window. On Pro, 1M context requires usage credits — which I never opted into and was never notified about. My existing, previously-working workflow stopped overnight with no warning.

Steps to Reproduce

  1. Be on the Claude Pro plan with usage credits disabled
  2. Open Claude Code (latest version)
  3. Send any prompt — even a trivial one
  4. Result: immediately blocked with the Usage credits required for 1M context error

Workarounds Tried — NONE Worked

I tried every documented and community-suggested workaround. All failed with the identical error:

  1. ❌ Added { "env": { "CLAUDE_CODE_DISABLE_1M_CONTEXT": "1" } } to ~/.claude/settings.json, restarted — still blocked
  2. ❌ Set CLAUDE_CODE_DISABLE_1M_CONTEXT=1 as a shell environment variable — still blocked
  3. ❌ Used --model flag to force a standard-context model — still blocked
  4. ❌ Restarted Claude Code and the terminal multiple times — still blocked
  5. ❌ Verified context with /context — cannot, every command is blocked

The error appears to fire before the model flag or env variable is evaluated, so there is currently no user-side way to recover. Enabling usage credits should not be the only escape hatch for a Pro user who never requested 1M context.

Expected Behavior

On the Pro plan, Claude Code should:

  • Default to the standard 200K context window
  • Work normally whenever the plan's included quota is available
  • Not require usage credits for standard operation
  • Honor CLAUDE_CODE_DISABLE_1M_CONTEXT=1 and --model flags

Actual Behavior

Claude Code silently selects a 1M-context model and hard-blocks all input at the auth layer, ignoring every override flag, demanding usage credits despite available quota.

Environment

  • Plan: Claude Pro
  • OS: Windows 11
  • Claude Code version: [FILL IN — run /status or check the title bar]
  • Interface: [CMD / PowerShell / VS Code terminal — FILL IN]
  • Model: default (never manually changed)
  • Usage credits: disabled (intentionally — should not be required)

Related Issues

Same root cause as: #62063, #62199, #63060, #63896, #63908, #64057

Requests

  1. Fix the regression: Pro users must not be gated behind usage credits for standard operation.
  2. Make overrides actually work: CLAUDE_CODE_DISABLE_1M_CONTEXT=1 and --model must take effect before the auth-level check, since right now neither does.
  3. Provide an immediate recovery path for users currently locked out with available quota.
  4. Notify Pro users in advance before introducing changes that require additional payment.
  5. Review and refund any usage unfairly consumed due to this bug.

What Should Happen?

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been adequately resolved
  • [x] This is a single bug report
  • [x] I am using the latest version of Claude Code

Bug Description

Claude Code on the Pro plan blocks every single prompt with the following error, and none of the documented workarounds resolve it:

API Error: Usage credits required for 1M context · turn on usage credits at claude.ai/settings/usage, or use --model to switch to standard context

The error fires at the authentication level, before any model flag or environment variable is processed, making Claude Code completely unusable on the Pro plan. This is a hard blocker.

This Is Not a Quota Issue

At the moment the error occurred, my Account & Usage panel showed plenty of remaining quota:

| Limit | Used | Remaining |
|-------|------|-----------|
| Session (5hr) | 17% | 83% |
| Weekly (7 day) | 15% | 85% |
| Weekly Sonnet | 14% | 86% |

Plan: Claude Pro

Every limit is far from exhausted, yet Claude Code is fully blocked. This is a feature/billing gate firing incorrectly, not a genuine usage limit.

This Worked Before

This is a regression. It began after a recent Claude Code update silently changed the default model to one using a 1M-context window. On Pro, 1M context requires usage credits — which I never opted into and was never notified about. My existing, previously-working workflow stopped overnight with no warning.

Steps to Reproduce

  1. Be on the Claude Pro plan with usage credits disabled
  2. Open Claude Code (latest version)
  3. Send any prompt — even a trivial one
  4. Result: immediately blocked with the Usage credits required for 1M context error

Workarounds Tried — NONE Worked

I tried every documented and community-suggested workaround. All failed with the identical error:

  1. ❌ Added { "env": { "CLAUDE_CODE_DISABLE_1M_CONTEXT": "1" } } to ~/.claude/settings.json, restarted — still blocked
  2. ❌ Set CLAUDE_CODE_DISABLE_1M_CONTEXT=1 as a shell environment variable — still blocked
  3. ❌ Used --model flag to force a standard-context model — still blocked
  4. ❌ Restarted Claude Code and the terminal multiple times — still blocked
  5. ❌ Verified context with /context — cannot, every command is blocked

The error appears to fire before the model flag or env variable is evaluated, so there is currently no user-side way to recover. Enabling usage credits should not be the only escape hatch for a Pro user who never requested 1M context.

Expected Behavior

On the Pro plan, Claude Code should:

  • Default to the standard 200K context window
  • Work normally whenever the plan's included quota is available
  • Not require usage credits for standard operation
  • Honor CLAUDE_CODE_DISABLE_1M_CONTEXT=1 and --model flags

Actual Behavior

Claude Code silently selects a 1M-context model and hard-blocks all input at the auth layer, ignoring every override flag, demanding usage credits despite available quota.

Environment

  • Plan: Claude Pro
  • OS: Windows 11
  • Claude Code version: [FILL IN — run /status or check the title bar]
  • Interface: [CMD / PowerShell / VS Code terminal — FILL IN]
  • Model: default (never manually changed)
  • Usage credits: disabled (intentionally — should not be required)

Related Issues

Same root cause as: #62063, #62199, #63060, #63896, #63908, #64057

Requests

  1. Fix the regression: Pro users must not be gated behind usage credits for standard operation.
  2. Make overrides actually work: CLAUDE_CODE_DISABLE_1M_CONTEXT=1 and --model must take effect before the auth-level check, since right now neither does.
  3. Provide an immediate recovery path for users currently locked out with available quota.
  4. Notify Pro users in advance before introducing changes that require additional payment.
  5. Review and refund any usage unfairly consumed due to this bug.

Error Messages/Logs

## Preflight Checklist

- [x] I have searched existing issues and this hasn't been adequately resolved
- [x] This is a single bug report
- [x] I am using the latest version of Claude Code

## Bug Description

Claude Code on the **Pro plan** blocks **every single prompt** with the following error, and **none of the documented workarounds resolve it**:


API Error: Usage credits required for 1M context · turn on usage credits at claude.ai/settings/usage, or use --model to switch to standard context


The error fires at the **authentication level, before any model flag or environment variable is processed**, making Claude Code completely unusable on the Pro plan. This is a hard blocker.

## This Is Not a Quota Issue

At the moment the error occurred, my Account & Usage panel showed plenty of remaining quota:

| Limit | Used | Remaining |
|-------|------|-----------|
| Session (5hr) | 17% | 83% |
| Weekly (7 day) | 15% | 85% |
| Weekly Sonnet | 14% | 86% |

Plan: **Claude Pro**

Every limit is far from exhausted, yet Claude Code is fully blocked. This is a feature/billing gate firing incorrectly, not a genuine usage limit.

## This Worked Before

This is a regression. It began after a recent Claude Code update silently changed the default model to one using a 1M-context window. On Pro, 1M context requires usage credits — which I never opted into and was never notified about. My existing, previously-working workflow stopped overnight with no warning.

## Steps to Reproduce

1. Be on the **Claude Pro** plan with usage credits **disabled**
2. Open Claude Code (latest version)
3. Send any prompt — even a trivial one
4. Result: immediately blocked with the `Usage credits required for 1M context` error

## Workarounds Tried — NONE Worked

I tried every documented and community-suggested workaround. **All failed with the identical error:**

1. ❌ Added `{ "env": { "CLAUDE_CODE_DISABLE_1M_CONTEXT": "1" } }` to `~/.claude/settings.json`, restarted — still blocked
2. ❌ Set `CLAUDE_CODE_DISABLE_1M_CONTEXT=1` as a shell environment variable — still blocked
3. ❌ Used `--model` flag to force a standard-context model — still blocked
4. ❌ Restarted Claude Code and the terminal multiple times — still blocked
5. ❌ Verified context with `/context` — cannot, every command is blocked

The error appears to fire **before** the model flag or env variable is evaluated, so there is currently **no user-side way to recover**. Enabling usage credits should not be the only escape hatch for a Pro user who never requested 1M context.

## Expected Behavior

On the Pro plan, Claude Code should:
- Default to the standard **200K** context window
- Work normally whenever the plan's included quota is available
- **Not** require usage credits for standard operation
- Honor `CLAUDE_CODE_DISABLE_1M_CONTEXT=1` and `--model` flags

## Actual Behavior

Claude Code silently selects a 1M-context model and hard-blocks all input at the auth layer, ignoring every override flag, demanding usage credits despite available quota.

## Environment

- **Plan:** Claude Pro
- **OS:** Windows 11
- **Claude Code version:** [FILL IN — run `/status` or check the title bar]
- **Interface:** [CMD / PowerShell / VS Code terminal — FILL IN]
- **Model:** default (never manually changed)
- **Usage credits:** disabled (intentionally — should not be required)

## Related Issues

Same root cause as: #62063, #62199, #63060, #63896, #63908, #64057

## Requests

1. **Fix the regression:** Pro users must not be gated behind usage credits for standard operation.
2. **Make overrides actually work:** `CLAUDE_CODE_DISABLE_1M_CONTEXT=1` and `--model` must take effect before the auth-level check, since right now neither does.
3. **Provide an immediate recovery path** for users currently locked out with available quota.
4. **Notify Pro users in advance** before introducing changes that require additional payment.
5. **Review and refund** any usage unfairly consumed due to this bug.

Steps to Reproduce

<img width="389" height="175" alt="Image" src="https://github.com/user-attachments/assets/d6211e57-fff8-4d92-9fab-56af257913c9" />

Claude Model

Sonnet (default)

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

sonnet 4.6

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_

View original on GitHub ↗

This issue has 6 comments on GitHub. Read the full discussion on GitHub ↗