Important Pricing Update: Reasoning Models & Deep Research
We know that not all news is good news, but we need to share this update with you to maintain our transparency and sustainability.
Why This Change is Necessary
At Straico, prices have been calculated using only input and output tokens, which doesn’t fully reflect our costs. This is because reasoning and citation tokens are also paid on our end, but we have never charged our users for them until now.
After a financial analysis, we’ve decided to increase the prices of all reasoning models and the Sonar Deep Research model.
Pricing Changes Overview
Reasoning Models
20% price increase across all reasoning models
Deep Research
12× price multiplier for Perplexity Sonar Deep Research
Updated Model Pricing
Reasoning Models (20% increase)
| Model | Previous Price | New Price |
| DeepSeek R1 | 4.3 coins | 5.16 coins |
| DeepSeek R1: Nitro | 4.3 coins | 5.16 coins |
| Claude 3.7 Sonnet (Reasoning) | 5 coins | 6.00 coins |
| Perplexity Sonar (Reasoning) | 2 coins | 2.40 coins |
| Qwen3 235B A22B | 0.5 coins | 0.60 coins |
| Gemini Flash 2.5 (Reasoning) | 1 coin | 1.20 coins |
| Anthropic: Claude Sonnet 4 | 5 coins | 6.00 coins |
Deep Research Model
| Model | Previous Price | New Price |
| Perplexity Sonar (Deep Research) | 16 coins | 192 coins |
Understanding the Deep Research Pricing
We understand that the price increase, especially for Deep Research, is significant. However, it reflects:
- The power of the tool and its comprehensive capabilities
- The vast amount of data it pulls from the web
- The intensive reasoning applied to that information
Future Improvements
In the near future, we’ll change the pricing calculation for these models to separate the coins charged for input/output tokens from those charged for reasoning and citation tokens. This will provide greater transparency in our pricing structure.
Our Commitment
We’re truly sorry for these changes, but they are absolutely necessary to keep Straico sustainable. We remain committed to providing you with cutting-edge AI capabilities while maintaining transparency about our costs and pricing.
