Is there a flat-rate API for Cline?
Yes: set Cline's API Provider to OpenAI Compatible with the base URL of a flat-rate API, such as Standard Compute's https://api.stdcmpt.com/v1, your key and the model standardcompute. You pay $19, $39, $89 or $249 a month for a fixed compute budget across Claude, GPT-5.6, Gemini, DeepSeek, GLM, Qwen and Kimi, and at 100% new requests get HTTP 402 instead of a charge. Cline's own provider bills per token, and ClinePass ($9.99 a month) covers only open-weight models.
API Provider OpenAI Compatible
Base URL https://api.stdcmpt.com/v1
OpenAI Compatible API Key YOUR_API_KEY
Model ID standardcompute
# MODEL CONFIGURATION
Context Window Size 400000
Max Output Tokens 32000
Your other providers stay configured in Cline, so you can switch back for any task. Cline's cost display shows $0 unless you enter prices; the dashboard's Usage page shows the budget used.
Every way to pay for models in Cline
| Option | What you pay | What happens at the limit |
|---|---|---|
| Cline provider | Each model's per-token price, from prepaid Cline credits, across 100+ models | At a zero balance you top up to continue |
| ClinePass | $9.99 a month for 12 open-weight coding models such as GLM-5.3, Kimi K3 and DeepSeek V4 Pro. No Claude or GPT | 5-hour, weekly and monthly limits whose size Cline does not publish. At a limit Cline offers its pay-as-you-go billing |
| Your Anthropic or OpenAI key | Per token. Claude Sonnet 5.5: $2 input / $10 output per 1M tokens; Opus 5.5: $4 input / $20 output | Your account's rate limits and any spend limit you set there |
| Standard Compute | $19, $39, $89 or $249 a month for a $20, $41, $95 or $269 compute budget | New requests get HTTP 402 until the plan renews. No overage, no 5-hour window, no weekly cap |
Cline terms from ClinePass and Cline provider docs, Claude prices from Anthropic's pricing page, checked October 14, 2026.
What a month of Cline costs per token, and the plan that covers it
A median day with a coding agent costs $3.67 on Claude Sonnet 5.5, $7.34 on Opus 5.5 and $0.23 on GPT-6 Luna at list prices, with caching. 20 such days on Sonnet come to about $73.40, inside the Standard plan's $95 budget. A heavy day reaches $29.55 on Sonnet (90th percentile), so a hard week can use the budget up before the month ends.
| Model on every request | Median day | 20 days per token | Smallest plan that covers it |
|---|---|---|---|
| Claude Opus 5.5 | $7.34 | $146.80 | Pro, $249 |
| Claude Sonnet 5.5 | $3.67 | $73.40 | Standard, $89 |
| GPT-6 Luna | $0.23 | $4.60 | Starter, $19 |
Budgets: Starter $20, Economy $41, Standard $95, Pro $269. Day costs from the Agent Burn Index, priced at list rates. Cline was measured with the other coding agents; your own days will differ.
Plan and Act: one model for each mode
- Tick "Use different models for Plan and Act modes" and Cline keeps a separate OpenAI Compatible model, and a separate Reasoning Effort, for each mode. Use the same base URL and key in both.
- Plan mode does few, long turns of reasoning: put a frontier model there, such as
anthropic/claude-opus-5.5oropenai/gpt-5.6-sol. - Act mode does the many edit and command turns, where most of the tokens go:
anthropic/claude-sonnet-5.5for hard code, orz-ai/glm-5.3-flash,deepseek/deepseek-v4.1-flashoropenai/gpt-5.6-lunato make the budget last. - Or leave both modes on
standardcompute: the router sends routine steps to an efficient model and hard ones to the frontier model you select on the Models page, and the plan price stays the same whichever it picks.
Set it up in Cline
- Pick a plan from $19 a month and create a key.
- In Cline's settings, set API Provider to OpenAI Compatible.
- Set Base URL to
https://api.stdcmpt.com/v1, paste your key, and enterstandardcomputeas the Model ID, or pin a model such asanthropic/claude-sonnet-5.5,anthropic/claude-opus-5.5orz-ai/glm-5.3-flash. - Under Model Configuration, raise Context Window Size (Cline uses 128K by default; 400000 for the router), and untick Supports Images if you pin a text-only model such as GLM 5.3.
- Set Reasoning Effort per mode (Cline's default is none). Cline's value overrides the Models page setting for every request.
- Run one small task and check the request in the dashboard's request log before a long one.
Rate limits and 429s on long Cline sessions
- Each key accepts 300 (Starter), 600 (Economy), 1,000 (Standard) or 2,000 (Pro) requests a minute. Above that the answer is HTTP 429 with Retry-After.
- Standard Compute has no GPUs of its own and reaches each model's providers through OpenRouter. A transient upstream 429 is retried up to twice, sometimes on another provider for the same model, before Cline sees an error.
- The main gateway is in Finland. From the Americas you can use the US endpoint,
https://us.stdcmpt.com/v1, with the same key.
The catch
The budget is a real cap. Pin Opus 5.5 to every Act step and a heavy month uses it up; new requests then get HTTP 402 (budget_exhausted) until the plan renews, unless you upgrade or turn on Pay as you go (Economy and up, usage plus 15%). If you already spend about $89 a month on Sonnet through Anthropic and want Sonnet on every step, a plan that covers it costs about the same: you buy a fixed bill, not a discount. And if open models within ClinePass's windows do your work, ClinePass at $9.99 costs less than any of our plans.
Where Standard Compute fits
Standard Compute is an OpenAI- and Anthropic-compatible API on monthly plans of $19, $39, $89 or $249. Cline connects as an OpenAI Compatible provider, and the same key works in Kilo Code, ZooCode, Claude Code, Codex and OpenCode. Every plan has the same model catalog, the dashboard shows the budget live, and nothing is billed on top unless you turn on Pay as you go.
FAQ
Is anyone running Claude Sonnet on a flat rate in Cline? Going direct to Anthropic costs me $80 to $120 a month, and the swings bother me more than the amount.
Yes, with a flat-rate API as Cline's provider. Anthropic's API bills Sonnet 5.5 per token ($2 input / $10 output per 1M tokens) and has no monthly API plan. On Standard Compute the month costs the plan price. Pin anthropic/claude-sonnet-5.5 in Cline and Pro ($249, $269 budget) covers $120 of Sonnet with room to spare; use the standardcompute model with Sonnet selected on the Models page and the router takes the routine steps, so Standard ($89, $95 budget) covers most of those months. If you want Sonnet on every step and spend about $100, Anthropic direct costs less than Pro: what you buy here is a fixed bill.
Opus fixed a bug in 20 minutes that I'd been staring at for two days. Worth every cent, but it was a lot of cents. How do people use top models every day without the bill exploding?
They keep the frontier model for the hard part. In Cline, put Opus 5.5 in Plan mode and an efficient model in Act mode, or leave the model on standardcompute and select Opus 5.5 together with an efficient model on the Models page: the router picks per request. At list prices a median agent day costs $7.34 on Opus 5.5 and $0.23 on GPT-6 Luna. On a flat plan from $19, the month costs the plan price; if the budget runs out, requests stop until renewal instead of adding charges.
Can I use GLM, Qwen or Kimi in Cline without opening an account with each provider? Ideally all through one fixed-price API.
Yes. Add Standard Compute as Cline's OpenAI Compatible provider (base URL https://api.stdcmpt.com/v1, your key) and enter z-ai/glm-5.3, z-ai/glm-5.3-flash, qwen/qwen3.8-max-0902 or moonshotai/kimi-k3 as the Model ID. They run on the same key as Claude and GPT, at a fixed monthly price from $19 with a fixed compute budget. ClinePass is Cline's own plan for a set of open models with time windows; Standard Compute covers the whole catalog and also works in Claude Code, Codex and OpenCode.
Is ClinePass a flat rate for Claude?
No. ClinePass costs $9.99 a month and covers 12 open-weight models, such as GLM-5.3, Kimi K3, DeepSeek V4 Pro and Qwen3.8 Max; Claude and GPT are not included. Its 5-hour, weekly and monthly limits have no published size, and at a limit Cline offers its pay-as-you-go billing. Claude in Cline is per token on the Cline provider or your Anthropic key, or a fixed monthly price through a flat-rate API.
What happens in Cline when the monthly budget runs out?
New requests get HTTP 402 with the code budget_exhausted, and Cline shows the error. The x-sc-budget-renews-at response header says when the plan renews. Nothing is charged on top: to keep working before then, upgrade (it takes effect immediately) or, on Economy and above, turn on Pay as you go, which adds compute in prepaid steps at the usage rate plus 15%.
Summary of my month: four OpenRouter top-ups, two DeepSeek top-ups and a card blocked for repeated Anthropic charges. This isn't sustainable. Who does one fixed monthly bill?
An API that puts those models behind one key and charges one fixed monthly price. Standard Compute serves Claude, DeepSeek, GPT-5.6, GLM, Qwen and Kimi on the same key, as Cline's OpenAI Compatible provider, for $19, $39, $89 or $249 a month: one Stripe charge, no top-ups. If the budget runs out, requests stop until renewal instead of charging more.
OpenRouter rate-limits me at peak hours, which is exactly when I work. Is there a provider with its own capacity and a fixed monthly price?
Own capacity, no: Standard Compute has no GPUs and also reaches providers through OpenRouter, so a general OpenRouter outage affects us too. What differs is the handling: a transient 429 is retried up to twice, sometimes on another provider for the same model, before your agent sees it; each key accepts 300 to 2,000 requests a minute by plan; and the price is fixed per month. If you need guaranteed capacity, what you want is reserved throughput from an inference provider, contracted separately.
I'm a student. Free Gemini was enough until I hit the daily cap, and now Google wants me to set up billing in GCP. I just want one key and an affordable monthly price. Options?
Standard Compute's Starter plan: $19 a month for a $20 budget, one key and no Google Cloud project. It includes Gemini 3.8 Flash and Gemini 3.5 Flash-Lite, and very cheap models such as DeepSeek V4 Flash ($0.14 input / $0.28 output per 1M tokens) that stretch the budget. Eligible new accounts can activate $0.25 of trial compute with no card to check that Cline works before paying.
Why does Cline show $0.00 for every request on my custom provider?
An OpenAI Compatible model shows $0 in Cline's cost display until you enter its input and output prices under Model Configuration. Billing is unaffected: on Standard Compute every request draws on the monthly budget, the Usage page shows how much is used, and the request log shows which model answered each call.