Confused about pricing

[unkempt shadow]

Hello, maybe I'm fundamentally misunderstanding something. I recently generated something using deepseek 3.1 and it shows that I've used 53k input and 370 output. Which by the math of $0.25/M in and $1/M out should mean about 1.36 cents? But somehow I was charged basically an extra cent at 2.44 cents. Can someone explain to me why that might be?

[austere apex]

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

To help clarify your pricing question, I need a bit more information:

• Can you provide the exact error message or billing line you see? (This helps identify the precise charge calculation.)
• Are you sure the model you used is exactly deepseek-3.1 (please specify the exact model identifier)?
• Was this a single request or multiple requests? If multiple, were they all for the same model?

With these details, I can help you (or escalate to support/mods if needed) to pinpoint the charge.

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

[unkempt shadow]

Model DeepSeek V3.1
Model ID deepseek/deepseek-chat-v3.1
Provider AtlasCloud
First token latency 5.56 seconds
Throughput 18.3 tokens/second
Finish reason stop
Data policy No data training | Policy
Tokens 53056 prompt 370 completion
Subtotal 0.0244
File processing cost 0
Final cost 0.0244

This was through all deepseek 3.1

[austere apex]

Thanks for the details. Here’s how the pricing works for DeepSeek models on OpenRouter:

• For DeepSeek V3 models (such as deepseek/deepseek-chat-v3-0324 and deepseek/deepseek-chat), the price is "$0.18 / M tokens input and $0.72 / M tokens output" Model Pricing.

Your usage:

• Input: 53,056 tokens
• Output: 370 tokens

Cost calculation:

• Input: 53,056 / 1,000,000 × $0.18 ≈ $0.00955
• Output: 370 / 1,000,000 × $0.72 ≈ $0.000266
• Total: ≈ $0.00981 (input) + $0.000266 (output) ≈ $0.01008

However, your charge is $0.0244, which is higher. A few possible reasons:

• The model you used is deepseek/deepseek-chat-v3.1 (provider: AtlasCloud), which is not listed in the provided DeepSeek V3 pricing docs. Pricing may differ for different providers or specific model versions, and not all DeepSeek models have the same pricing.
• Rounding, minimum charge policies, or provider-specific fees may apply.

"I don't have specific documentation for deepseek/deepseek-chat-v3.1 pricing. Please verify with the mods or check the OpenRouter docs directly for the most accurate information."

For billing disputes or exact charge clarifications, email [e-mail], as billing issues are not handled on Discord.

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

[unkempt shadow]

I also had similar charges where they were close to 3 cents

[zenith comet]

every provider has different prices, check it on https://openrouter.ai/deepseek/deepseek-chat-v3.1

[austere apex]

Thank you for clarifying that you used "deepseek 3.1." There are multiple DeepSeek models with different pricing, so it's important to have the exact model identifier to determine the correct rate. For example, models like deepseek-chat-v3-0324, deepseek-v3-base, or deepseek-chat may have different token prices.

From the DeepSeek model listings:

• deepseek-chat-v3-0324: $0.20/M input, $0.80/M output
• deepseek-v3-base: $0.18/M input, $0.72/M output
• deepseek-chat: $0.18/M input, $0.72/M output

If you used a different DeepSeek model (such as R1, R1-Zero, or distilled variants), the pricing can be significantly higher or lower. For example, DeepSeek R1 is $0.40/M input, $2/M output, and some distilled variants are as low as $0.04/M input/output.

Since the total charge doesn't match your expected math, it's possible that:

• You used a DeepSeek model with different pricing than you thought
• There were multiple requests or rounding differences
• Output tokens were rounded up, or there was an extra charge for minimum usage

Please double-check the exact model identifier you used in your requests. You can compare the exact rates for each model at the DeepSeek | OpenRouter page.

If your charges still seem off after checking the model and rate, you should email [e-mail] with your account details and billing line for a detailed review. Billing issues are not handled on Discord.

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

[unkempt shadow]

Oh.. so its not just a one size fit all because on OR it shows .25 per million

[zenith comet]

that's the price of the cheapest provider, which is Chutes

[unkempt shadow]

Okay, that makes a lot more sense now.

Thanks Sycho! I'll put some of the more higher pricing ones on ignore list then. I thought it was the same for all hah

[zenith comet]

either that or add the cheaper ones to approved providers 🙂

[unkempt shadow]

So does approved means it will only use those? Does it work on a white list or black list priority?

[zenith comet]

approved is a whitelist (you might want to enable the 'Always enforce'), ignored is a blacklist.