anthropic/claude-sonnet-4.5 the 200k was allowed to use more than 200k tokens and was billed extra

[serene smelt]

I'm an AI assistant - my answers may be incorrect and I cannot see screenshots. Please verify information with the OpenRouter docs.

Thanks for clearly describing your issue. To help escalate and diagnose this unexpected billing behavior, I need these details:

1. What was the exact API request you made (including which endpoint and parameters)?
2. What was the full error or response (if any) when you sent a context over 200k tokens to anthropic/claude-sonnet-4.5?
3. Can you confirm the exact model identifier used (e.g., anthropic/claude-sonnet-4.5 — and not one of the 1M context variants)?

Once you provide this info, I can help the support team investigate why you were billed for a 1M context request when you expected 200k only.

Documentation shows, "Claude Sonnet 4.5: context length 1M" and pricing tables for ≤200K and >200K token usage, but does not mention automatic switching or forced billing at higher tiers. Always confirm your usage and limits in your API parameters for safety Claude Sonnet 4.5 - API, Providers, Stats.

For billing/account disputes, you must contact [email protected]. Discord does not handle billing issues.

-# This is an automated response powered by Answer HQ. Use !mute to stop the bot from responding.

Thank you for these details. Based on the OpenRouter documentation for anthropic/claude-sonnet-4.5, the context length can go up to 1 million tokens, and the pricing table shows tiered rates:

• "| 1M | 64K | ≤200K $3 >200K $6" for input
• There is no mention of automatic context restriction; requests over 200K tokens are billed at a higher rate, not rejected or limited Claude Sonnet 4.5 - API, Providers, Stats.

So, if you send a prompt over 200K tokens to anthropic/claude-sonnet-4.5, you’re charged at the higher >200K rate, but it does not switch you to a separate 1M model. This is by design, not a bug. The charge you saw (e.g., $5+) is consistent with usage above 200K tokens at the posted rates.

"| 1M | 64K | ≤200K$3>200K$6 |" (input), which means any input over 200K tokens is charged at the higher rate for this same model, not a different one Claude Sonnet 4.5 - API, Providers, Stats.

If you want to avoid higher charges, you must ensure your input size never exceeds 200K tokens on your end, as OpenRouter does not cap or error at the boundary.

For billing or refund questions, contact [email protected].

Full pricing and context policy found at Claude Sonnet 4.5 - API, Providers, Stats.

-# This is an automated response powered by Answer HQ. Use !mute to stop the bot from responding.

[drifting dove]

there's no 200k variant,

Sonnet 4.5 by default is 1M, and the higher pricing applies when the context surpasses 200k tokens

:1M does not exist as a variant in their documentation

"their" as in OpenRouter

i do not work for OpenRouter

and no it doesn't exist

https://openrouter.ai/anthropic/claude-sonnet-4.5

you can see all providers offer 1M tokens of context

that's a Cline mistake then

if you call Anthropic directly, they offer 200k context and 1M if you use their beta heads

the Cline team might have applied that for OpenRouter directly unknowningly that all providers are set to 1M by default on OpenRouter

if you see in the Cline UI, it's probably by Cline.

if you call the /models enpoint by OpenRouter, you'll see theres no 200k variant

[solar iris]

cairo is correct, we don't have a 200k context model

there is only one model, and it has 1M

https://openrouter.ai/anthropic/claude-sonnet-4.5

[solar iris]

that's not an openrouter thing

doesn't exist on openrouter