AI usage limits and API errors, explained
How many messages a plan really gets, and what the error on your screen means. Each answer was checked against first-party pages on the date it shows.
Does Max effort on Claude Opus 5.5 use up your limit faster?
Thinking can't be switched off on Opus 5.5, so the effort level is the only setting left that changes how fast your limit goes.
How do you use the free Claude limit reset that came with Opus 5.5?
It's in Settings > Usage on the web or in Claude Desktop, it can't be undone, and it expires if you sit on it.
“You exceeded your current quota, please check your plan and billing details.” (OpenAI API)
Why the OpenAI API says “You exceeded your current quota, please check your plan and billing details.”, the four billing codes behind insufficient_quota, why retrying doesn't help, and how Codex shows the same refusal.
“Error: pull model manifest: file does not exist” in Ollama: the name or tag isn't on the registry
What Ollama means by “Error: pull model manifest: file does not exist”, why it's really a 404 from the model registry, the :cloud suggestion newer versions add, and the /save bug that prints it for a model you already have.
“Usage limit reached for all Pro models.” in Gemini CLI: what ran out and what to do
What Gemini CLI means by “Usage limit reached for all Pro models.”, how it differs from the “high demand” and “exhausted your capacity” messages, the daily request limits per plan, and what to try when /stats says you have quota left.
“429 RESOURCE_EXHAUSTED” from the Gemini API: which quota you hit and how to tell
What the Gemini API's “429 RESOURCE_EXHAUSTED” and “You exceeded your current quota” mean, how to read the metric line under them, what limit: 0 tells you, and when retrying is pointless.
“You've reached your weekly rate limit.” in GitHub Copilot CLI: waiting it out or switching to auto
What GitHub Copilot CLI means by “You've reached your weekly rate limit.”, how it differs from the session and per-model limits and from running out of AI credits, and the setting that switches to auto for you.
“You’ve hit your usage limit.” in Codex: what each version of the message means
Why Codex says “You’ve hit your usage limit.”, why the rest of the line changes with your ChatGPT plan, how to read the reset time, and what to do when /status says you have usage left.
“stream disconnected before completion” in Codex: what drops the stream and what to try
What Codex means by “stream disconnected before completion”, how its automatic reconnects work, the Transport error and request ID variants people paste, and what has fixed it for them.
“Selected model is at capacity. Please try a different model.” in Codex, explained
Why Codex says “Selected model is at capacity. Please try a different model.”, what OpenAI staff say it means, why Codex doesn't retry it for you, and what to do when every model says it.
Codex models and plans explained: GPT-6 Astra, Sol and Luna, and which plan gets which
Pick any GPT-6 model you like on your laptop, but a Codex cloud task on a ChatGPT plan still runs on last generation's Sol.
“Credit balance is too low” in Claude Code and the Claude API: what it checks and how to clear it
Why Claude Code or the Claude API says “Credit balance is too low”, including on a Pro or Max plan, and what cleared it for people who had already topped up.
Claude Code vs Codex vs Gemini CLI: which plan you need and which models you get
One of the three stopped serving individual subscribers in June, but its own docs still advertise the free tier it closed.
“Prompt is too long” in Claude Code: why it happens and how to get the session back
What Claude Code's “Prompt is too long” and “Context limit reached” mean, why /compact sometimes can't rescue the session, and how MCP tools and big pastes fill the window before you've typed much.
“OAuth token has expired. Please obtain a new token or refresh your existing token.” in Claude Code
Why Claude Code shows “API Error: 401 … OAuth token has expired”, why the refresh that should prevent it failed, and what to do when /login doesn't stick.
“Not logged in · Please run /login” in Claude Code: why /login doesn't always stick
What Claude Code's “Not logged in · Please run /login” and “Invalid API key” lines mean, how Claude Code picks a credential, and why the prompt comes back after /login says it worked.
“API Error: Request timed out.” in Claude Code: which timeout fired and what to change
What Claude Code means by “API Error: Request timed out.”, how it differs from the network version and from “No response from API”, and which setting to raise for a slow proxy.
“API Error: Connection dropped (ECONNRESET)” in Claude Code, and the older “Connection error.”
What Claude Code means by “API Error: Connection dropped (ECONNRESET)” and “API Error: Connection error.”, what the official error reference says to check, and why a VPN fixed it for some people.
“API Error: 400 due to tool use concurrency issues. Run /rewind to recover the conversation.” in Claude Code
What “API Error: 400 due to tool use concurrency issues.” means in Claude Code, the raw “tool_use ids were found without tool_result blocks” error behind it, and when /rewind won't fix it.
What are the GPT-6 Sol and Luna rate limits on the OpenAI API?
Sol's table is Astra's table, number for number. Luna's is the one that's different, and at Tier 5 it carries four and a half times the tokens a minute.
What are the GPT-6 Astra rate limits on the OpenAI API?
The free tier can't call it at all, and the Tier 2 to Tier 3 step leaves requests per minute exactly where they were.
“/limit-reset isn't available in this environment.” Why Claude Code says it
Why Claude Code answers /limit-reset with “/limit-reset isn't available in this environment.”, which sessions print it, what it doesn't mean for your account, and how to run the reset from a terminal session instead.
Claude API invalid signature in thinking block: what it means and how to fix it
What the Claude API's invalid signature in a thinking block means, why editing an earlier message or the system prompt causes it, which accounts the prefix check runs on by default, and the two fixes.
“A session-limit reset isn't available right now.” What Claude Code's /limit-reset means
Why Claude Code's /limit-reset prints “A session-limit reset isn't available right now.”, the checks that have to pass before a reset is offered, the reasons the server can hold back without telling you, and the other lines the command prints.
What do you actually get on Claude Max 5x and Max 20x?
The multiplier is on the session allowance, and a separate weekly ceiling sits over every model you use.
How many GPT-6 Astra messages do you get on ChatGPT Pro $100 and Pro $200?
Pro $200 keeps GPT-6 Pro on its own weekly counter; Pro $100 makes it share one allowance with GPT-5.6 Sol Pro.
How do you keep using Claude after Fable 5.1 eats your plan's usage?
The Fable half of a Max plan's weekly limit comes out of that limit, so running past it means paying API rates for the rest.
What happened to the GPT-5.3-Codex-Spark usage limit on ChatGPT Pro?
The separate quota was tied to a research preview, and the preview is over.
What happens when you hit your GPT-6 Astra limit on ChatGPT Plus or Pro?
Credits pick up where the plan stops, they're good for a year, and the balance can end up below zero.
llama.cpp unknown model architecture: 'qwen4exp': what it means and how to fix it
What llama.cpp means by unknown model architecture: 'qwen4exp', which pull request added Qwen3.8-Flash-Next support and when, the Ollama release that carries it, and the load failures that look similar.
What's the context limit on Qwen3.8-27B, and how do you push it to 1M?
The million-token number needs a config change, and Qwen says leaving that change on hurts short prompts.
Is Muse Spark 1.3 free on OpenCode, and what do you get for $10 a month?
The free tier is a data trade, and the paid one is metered in dollars rather than messages.
llama.cpp hangs loading Qwen3.8-Flash-Next on ROCm: what the last log line means
Why llama-server stops partway through loading Qwen3.8-Flash-Next on llama.cpp's HIP backend, why the hang isn't an out-of-memory failure, what sets it off, and what loads the same file instead.
How many credits does GPT-6 Astra cost in ChatGPT?
The chat side charges by the message and the Codex side charges by the token, and Fast mode multiplies the token side by 2.5.
How much Claude Opus 5 do you get on Pro and Max in the Claude apps?
Anthropic meters Opus by how hard the work is, so the help center gives you five-hour sessions and a weekly Opus limit instead of a message count.
Is DeepSeek V4 Flash free in Cline, and what do you get for $9.99?
Cline's own pages put no token count, no request count and no reset window against the ClinePass quota.
How much Claude Opus 5 can you run in Claude Code on Pro and Max?
Max 5x and Max 20x are sold as multiples of the Pro allowance, and the help center leaves the Pro figure out.
How much Muse Spark 1.3 usage do you get on Cursor?
Cursor bills it out of the Other Models pool at Meta's list price, so what you get depends on how you use it.
Claude API 429 rate_limit_error: why Fable 5.1 hits it before Opus 5 does
What a 429 rate_limit_error from the Claude API means, how to tell a rate limit from the monthly spend cap, why Claude Fable 5.1 runs into it sooner than Opus 5 or Sonnet 5, and what counts toward the token limits.
Claude API 529 overloaded_error: what it means, and why it can arrive after a 200
What a 529 overloaded_error from the Claude API means, why it is not about your usage, how it arrives inside a streaming response after a 200, and what the SDKs already do about it.
What does the DeepSeek V4 Pro API cost, and what are its rate limits?
Peak covers seven hours a day, Monday to Friday, so most of the week bills at half the peak rate.
How many DeepSeek V4 Flash tokens do you get on Ollama Cloud Pro?
Ollama doesn't publish a token allowance for the plan, so the answer is $60 divided by a per-million rate.
Gemini PROHIBITED_CONTENT: what the API means by it and what to do
What PROHIBITED_CONTENT means in the Gemini API, the difference between seeing it as a block reason and as a finish reason, why safety settings can't switch it off, and the labels that look like it.
Claude Fable 5.1 doesn't support forced tool use: what the 400 error means
Why Claude Fable 5.1 and Claude Mythos 5.1 return a 400 invalid_request_error for tool_choice any or tool, what Anthropic's docs say to send instead, and the other 400 errors that only these models produce.
How many GPT-6 Astra messages do you get on ChatGPT Plus and Pro?
The table gives a range, and the five-hour window is usually what stops you.
How much Claude Fable 5.1 usage do you get on the Pro and Max plans?
The split runs between Max and Pro, and the only number Anthropic publishes is a percentage.
“Lower-priority mode isn't available right now.” What Claude Code means
Why Claude Code prints “Lower-priority mode isn't available right now.”, what it rules out, the seven conditions that have to hold before /low-priority is offered at all, and the other messages the same feature can print.
Claude Code error 529 Overloaded: what it means and what to do
What a 529 overloaded_error is, why Claude Code says “Repeated 529 Overloaded errors”, how its retries work, why it offers to switch you from Opus to Sonnet, and the 429, 500 and 504 messages that look similar.
OpenAI paused the $200 ChatGPT tier. Per Astra message, it was the cheap one.
Two hundred dollars a month bought four times the Astra allowance of the hundred-dollar plan, which is the arithmetic behind the pause.
Claude Code's /limit-reset, and what it actually resets
A command that turns your five-hour session limit back on, once in a while, and tells you when you can do it again.