Cost and usage issues on Windows — page 2
Issues the maintainers labelled both platform:windows and area:cost.
548 issues · 171 open · 377 resolved (69%) · first seen Jul 26, 2025
Is this getting better or worse?
This class of problem is holding steady. 183 new reports in the last 90 days vs 223 in the 90 before — -18%. The open backlog peaked at 171 in 2026-10 and sits at 171 today.
Derived by counting each matching issue from its creation date until it closed. First-party
computation over the anthropics/claude-code tracker — the trend is not published
anywhere upstream.
Which builds report this
375 of these reports name the Claude Code build they were running, spanning 144 releases. Heaviest reporters:
- Claude Code v2.1.126
- Claude Code v2.1.87
- Claude Code v2.1.296
- Claude Code v2.1.119
- Claude Code v2.1.263
- Claude Code v2.1.247
- Claude Code v2.1.123
- Claude Code v2.1.260
Report volume, not a defect count: an issue can mention a build without that build having caused the failure. Versions come from the version string in the bug-report template, matched against real published tags.
Do these get fixed?
Median time to close is 37 days across 377 closures. Of the 377 closures with a recorded reason, 11% were closed as completed and 337 as not-planned or duplicate. The most recent completed fix landed in 2026-10.
Workarounds reported by the community
59 of these issues post a workaround someone says worked. The highest-engagement ones:
- [BUG] Extremely rapid token consumption. There appears to be a critical issue with token calculation and quota management. Users, including myself, are experiencing an unusually high rate of token consumption
- ….3% , meaning ~20x token inflation per turn. After applying workarounds it stabilized at 89-99% . Full analysis with per-request measured data, safe workarounds, and community references (including [cc-cache-fix](
Found in the thread of #38239 · still open - [BUG] Error: Error during compaction: 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
- …runs /clear. This is a new behavior as of a day or so ago. Workaround Running /clear restores the session. The error message suggests using --model to switch to standard context, but this requires hardcoding a model nam…
Found in the description of #63896 · still open - Claude Code changed default model to 1M context without notifying Pro users
- …overnight. I lost hours trying to debug the issue, applying workarounds (environment variables, settings.json), and ultimately had to downgrade to a smaller context model — losing functionality I was already using and pa…
Found in the description of #62199 · still open - Add --max-context flag to cap context window usage
- …access to current context token count, so can't build this as a workaround Environment - Claude Code on Windows (MSYS2/Git Bash) - Opus 4.6 with 1M context - Previously worked well with 200K limit
Found in the description of #34650 · resolved - Opus 1M context on Max plan requires extra usage despite docs saying 'included with subscription'
- …t yesterday. Upgraded from Max 5x → Max 20x today to try to work around this. Still rate-limited at 0% usage. Total spend now exceeds £300 across plan upgrades and erroneous extra-usage charges. /bug report submitted — F…
Found in the thread of #39841 · still open
Excerpts are quoted from the linked issue so you can judge relevance before opening it. Threads that only ask for a workaround, or report that there is none, are excluded. These are community suggestions, not official guidance from Anthropic — read the full thread before applying one.
Matching issues 101–200
Ranked by community engagement (reactions weighted over comments). Listing the top 300 of 548 matches; the statistics above cover all 548.
- [BUG] LIMIT calculation is broken - Weekly Limit reached, resetted 10pm -> 11pm weekly reached again
- [BUG] Billing between claude.ai and claud code is confusing.
- [Bug] Session token limit appears too restrictive for extended coding workflows
- [BUG] Rate limit blocks all commands including /status after upgrading to Max plan
- VSCode Extension: File modification reminders consume excessive context
- [Bug] Excessive token consumption in multi-turn conversations causing premature rate limit exhaustion
- [BUG] Claude Code CLI (`claude -p`) burns ~15-16k tokens of fixed overhead per call when used as a stateless batch text-completion backend
- Fable 5.1 weekly limit is never surfaced in the CLI: the client receives it, labels it 'Fable', then drops it
- Background agent keeps running and billing after its window is closed AND after logout/re-login — invisible from the active session, found only by luck on another device
- abrupt consumption of limits
- Token consumption during offline period — server-side session not terminated after user exit
- usage error on claude
- [BUG] CLAUDE_CODE_USE_BEDROCK=1 writes <2% of prompt to cache at >150k context, vs full prompt on Anthropic 1P direct
- [BUG] API Error: Extra usage is required for 1M context · run /extra-usage to enable, or /model to switch to standard context
- 3 free ultrareview runs consumed but crashed due to server-side rate limit — no findings returned
- You're out of extra usage · resets 7pm (Europe/Barcelona)
- [Bug] Unexpected high token consumption with non-Claude subagents and orchestrator models
- [Bug] Unexpected token consumption without user interaction or API calls
- [BUG] /ultrareview: three back-to-back runs produced no usable findings across different diff shapes (Dedupe crash → silent empty → silent empty)
- [BUG] Ultrareview crashes repeatedly on large PRs – no findings produced
- [Bug] Claude Code 4.7 High Token Usage and Context Limitations
- Usage limits + forced T&C changes + model removal = no real user choice
- [BUG] Usage limit reached despite UI showing low consumption percentages
- [BUG] Claude Code input unresponsive — "Usage limit reached" despite being well under 5-hour (28%) and weekly (16%) limits on Max 20x plan
- System prompt size tripled after update to v2.1.101, causing 3x faster quota consumption
- [BUG] incorrect API error 400
- [Bug] statusline-setup agent causes excessive token usage via nested subagents on Windows
- [BUG] Claude Code extension causing abnormal background requests leading to rapid rate limit exhaustion
- Context compaction drops subagent results, causing redundant re-runs and wasted tokens
- Background task output polling blocks /compact requests, consuming entire context window
- [Bug] Usage limit increases by 1% on startup without using token
- [BUG] Claude Code confirms stop but silently continues in the background spending tokens forever ("Tool permission request failed: Error: Stream closed")
- 5-hour limit reached ∙ resets 12pm
- [Bug] Fable 5.1 incorrectly blocked by usage credits despite included Max plan allowance
- [Bug] Rate limit consumption disproportionate to measured token usage for opus-5-1m with prompt caching
- Agent tool: uncontrolled recursive subagent spawning — 5 planned agents escalated to 361+ completed background agents, burned full 5h usage quota
- Claude Code requires 1M context credits for standard agentic tasks on Pro plan
- [BUG] Claude start session with 100% usage!
- [BUG] Unexplained token consumption spike since early April with no change in usage pattern
- [BUG] Error: Extra usage is required for 1M context · run /extra-usage to enable, or /model to switch to standard context
- Agent launches background sub-agents without user approval, wasting tokens
- [Bug] 72% of API requests are empty polling requests counting toward usage quota
- [BUG] Claude Code agent caused ~5 USD financial damage through repeated configuration errors
- [BUG] /fast toggle reports "extra usage credits exhausted" on Pro Max with plenty of plan quota remaining
- Excessive session usage: reading a single MEMORY.md file (~150 lines) consumes ~9% session limit on Sonnet
- [BUG] Billing: Pro subscription and $20 credits revoked without explanation — no Usage Policy violation identified
- [BUG] Weekly Limite Use Ghost
- Anthropic API Billing: Persistent Model Access Restriction
- [BUG] When using Claude Code my API credits are being consumed despite having a Pro subscription
- Claude AI usage limit reached|1753560000
- [Bug] Claude Code consumes credits and produces unwanted output despite user instructions
- [BUG] Desktop Code tab usage chart miscounts tokens: duplicate transcript lines summed, workflow agents skipped, 7d legend shows all-time
- [BUG] Unrecognized slash commands moved from a free, deterministic rejection to a paid, highly variable model call, with cost rising undocumented across releases
- [BUG] Image tool results can't be persisted, so budget eviction replaces them with a sentinel mid-history and invalidates the prompt cache
- Both Claude Desktop + standalone Claude Code CLI installed -> subscription usage silently billed to Extra Usage; only uninstalling the CLI fixes it (/login to subscription does not)
- [BUG] This is frustating to get usage limit issue because of Anthropic capacity limitation
- [BUG] Claude Code stuck with "Usage credits required for 1M context" even after /clear and available usage limit
- [BUG] Model selection not persisting - Max subscription triggers usage credits error for 1M context
- [BUG] Massive token/quota spike and Prompt Caching failure after 1M context hotfix
- [BUG] Claude Code v2.1.160 forces 1M context on Windows VS Code, blocks session continuation after error
- [BUG] Why I can not use Sonnet even I have enough quota?!
- [BUG] 5-hour limit triggers just by launching the CLI (zero messages sent)?
- [Claude Desktop] Background token consumption without user interaction — 5% weekly usage lost overnight
- [BUG] Session usage counter shows 38% used with no user activity after login
- VS Code Claude extension API usage limit reached
- C:/Program Files/Git/ultrareview crashed before producing findings — burned a usage
- [Bug] Infinite thinking loop consuming all tokens without tool execution
- [BUG] Claude code usage shows no usage while constant high usage
- [BUG] API Error / Usage exhausted when I just started and have plenty of usage left.
- [BUG] Claude Code generated automated claude-bridge subprocess that caused account suspension — no-code user, $367 lost
- [Bug] Usage spike after rate limit reset with resumed session
- [BUG] Inconsistency in the usage calculation system
- [BUG] Hook loop silently burns all context tokens — session dies with no warning that a hook caused it
- [BUG] Token counter under stats very off kilter input token 1/10th of output tokens, seems backwards
- [BUG] Consumed 100 pc session usage said request timed out.
- [BUG] Command /usage does not recognize subscription
- [Bug] Team member spawn fails intermittently with rate limit error
- [BUG] Weekly token burn rate increased significantly overnight - no local changes (Max 20x plan)
- [BUG] Unexpected high usage (49%) for minimal session work
- inconsistent token limit status and reset time
- context burning in Opus when switch to Sonnet is ignored
- [Bug] Rate Limits Severely Reduced for 5x Subscription Plan
- [BUG] Disable or limit the total token count of haiku automatically summarizing the previous session
- Model completely ignores explicit --ultrathink instruction and wastes tokens on unwanted sub-agent analysis
- [BUG] Fable weekly usage counted (20%) with zero Fable requests since reset — only Opus 5.5 / Sonnet 5 used (Max 20x, Windows)
- Default subagents spawn with ~60k tokens of inherited context; workflow subagents receive the user's latest chat line as their governing instruction
- You've hit your session limit · resets 2:40am (America/Bogota)
- [Bug] code-review consumes excessive token budget in short timeframe
- Fable used as the advisor does not appear on the Fable weekly usage meter (Max 5x)
- Background task kill/completion notification wakes the session and re-reads full context -- costs a full turn with no user input
- Fable 5.1 in Claude Code ignores a standing "delegate, don't code" instruction and burns premium usage on grunt work
- [BUG] A 13 hour 5-hour limit issue
- [BUG] skipWorkflowUsageWarning is undocumented and agent-writable — an agent can disable its own spend guardrail
- Desktop: ~3.9k tokens of built-in MCP tool schemas load non-deferred with no opt-out; built-in server instructions exceed the documented 2KB cap
- Malformed tool calls leaking into chat + disproportionate usage-limit consumption on Opus (regression, ~1 week)
- [BUG] Model chosen via /model silently reverts to default when session continues from another device (quota still billed to pinned model)
- [BUG] --resume causes massive silent token drain on session restore (Pro, v2.1.193)
- Agent ignores user prohibitions, wastes 96% weekly token limit with zero results
- Billing-header nonce (`cch`) in system prompt breaks prompt caching on third-party providers
- [BUG] Extreme quota consumption (here we go again)