| AI Spend by Model Family | Your spend on AI models grouped by the family of the model, such as Claude, GPT or Gemini, whether you buy them from the model maker, through Amazon Bedrock, through Google Cloud or through Cursor. Spend that is not tied to a model, such as seats, is shown on its own line. | AWS, GCP, OpenAI, Anthropic, Cursor | model family |
| All AI Spend by Provider | Your total AI spend in one place: the AI and machine learning services of AWS, Google Cloud and Azure, plus OpenAI, Anthropic and Cursor, grouped by provider so you can compare them. | AWS, GCP, Azure, OpenAI, Anthropic, Cursor | cloud provider |
| Anthropic - Cache Hit Rate by Model | The share of the input tokens each Claude model processed that came from the prompt cache instead of being read fresh, through the Anthropic API. A higher share means you pay the cheaper cache price for more of your input. Spend that is not billed per token is left out. | Anthropic | service |
| Anthropic - Cost by API Key Name | Your Anthropic spend grouped by API key, so you can see which keys cost the most. | Anthropic | API key |
| Anthropic - Cost by Model | Your Anthropic spend grouped by model, so you can see which models cost the most. | Anthropic | service |
| Anthropic - Cost by Token Type | Your Anthropic spend grouped by token type, so you can see how much goes to input, output, and cached tokens. | Anthropic | token type |
| Anthropic - Cost by Workspace Name | Your Anthropic spend grouped by workspace, so you can see which workspaces cost the most. | Anthropic | account |
| Anthropic - Price per 1M Tokens by Model | What each Claude model costs you per million tokens through the Anthropic API, cached tokens included. Spend that is not billed per token, such as web search, is left out. | Anthropic | service |
| Anthropic - Tokens by Model | How many tokens each Claude model processed through the Anthropic API, cached tokens included, so you can see where the volume goes. | Anthropic | service |
| Anthropic - Tokens by Workspace | How many tokens each Anthropic workspace processed, cached tokens included, so each team can see its share of the volume. | Anthropic | account |
| AWS - AI Cost | Your AWS AI and machine learning spend grouped by service, so you can see which AI services cost the most. | AWS | service |
| Bedrock - Cache Hit Rate by Model | The share of the input tokens each model processed on Amazon Bedrock that came from the prompt cache instead of being read fresh, including models you buy through AWS Marketplace. Models and lines that bill no cache are shown at zero or left out. | AWS | model |
| Bedrock - Cost by Inference Profile | Your Amazon Bedrock spend by the inference profile or model it was billed on, including regional and global cross-region profiles and models you buy through AWS Marketplace, so you can see which profiles cost the most. Spend with no profile is shown on its own line. | AWS | inference profile |
| Bedrock - Cost by Model | Your Amazon Bedrock spend by model, so you can see which models cost the most. Spend that is not billed per token, such as provisioned throughput, is shown on its own line. | AWS | model |
| Bedrock - Cost by Routing | Your Amazon Bedrock spend split between requests served in the region they were sent to and requests routed across regions, which bill at different rates. Spend that is not billed per token is shown on its own line. | AWS | routing |
| Bedrock - Cost by Service Tier | Your Amazon Bedrock spend by service tier: standard, priority, flex and batch. Spend that is not billed per token is shown on its own line. | AWS | service tier |
| Bedrock - Cost by Token Type | Your Amazon Bedrock spend by token type: input, output, cache read, cache write and embedding. Spend that is not billed per token is shown on its own line. | AWS | token type |
| Bedrock - Cost by Usage Type | Your Amazon Bedrock spend broken down by model and token type, so you can see which models and which input or output tokens cost the most. | AWS | usage type |
| Bedrock - Cost by Vendor | Your Amazon Bedrock spend by the company that makes the model, such as Anthropic, OpenAI, Meta or Amazon, including models you buy through AWS Marketplace. Spend that is not billed per token is shown on its own line. | AWS | vendor |
| Bedrock - Price per 1M Tokens by Model | What each model costs you per million tokens on Amazon Bedrock, including models you buy through AWS Marketplace. Spend that is not billed per token, such as provisioned throughput, is left out. | AWS | model |
| Bedrock - Price per 1M Tokens by Token Type | What input, output, cache read and cache write tokens cost you per million on Amazon Bedrock, including models you buy through AWS Marketplace, so you can see which kind of token drives the bill. | AWS | token type |
| Bedrock - Tokens by Model | How many tokens each model processed on Amazon Bedrock, including models you buy through AWS Marketplace, so you can see where the volume goes. | AWS | model |
| Bedrock - Tokens by Token Type | How many input, output, cache read and cache write tokens you processed on Amazon Bedrock, including models you buy through AWS Marketplace. | AWS | token type |
| Bedrock - Tokens by Vendor | How many tokens each model maker, such as Anthropic, OpenAI, Meta or Amazon, processed on Amazon Bedrock, including models you buy through AWS Marketplace. | AWS | vendor |
| Azure - AI Cost | Your Azure AI and machine learning spend grouped by service, so you can see which AI services cost the most. | Azure | service |
| Claude - Cost by Platform | Your spend on Claude models, split by where you buy them: directly from Anthropic, through Amazon Bedrock, through Google Cloud or through Cursor. | AWS, GCP, OpenAI, Anthropic, Cursor | cloud provider |
| Cross-CSP - Total AI Cost | Your total AI and machine learning spend across AWS, Google Cloud, and Azure, grouped by cloud so you can compare AI cost across providers. | AWS, GCP, Azure | cloud provider |
| Cursor - Active Users per Seat | Of the people you pay a Cursor seat for, the share who actually used it in each period. A falling line means seats are going unused. | Cursor | Not grouped (one Total group) |
| Cursor - Active Users per Seat by Billing Group | The share of paid Cursor seats in each billing group that were actually used, so you can see which teams have adopted it. | Cursor | billing group |
| Cursor - Cache Hit Rate by Model | The share of the input tokens each model processed in Cursor that came from the prompt cache instead of being read fresh. Seat fees are left out. | Cursor | service |
| Cursor - Cost by Billing Group | Your Cursor spend, seats included, by Cursor billing group, so each team can see and own its share. | Cursor | billing group |
| Cursor - Cost by Model | Your Cursor spend by model, with seat licenses as their own line, so you can see which models drive usage-based charges. | Cursor | service |
| Cursor - Cost by Request Source | Your Cursor spend by where requests come from: the editor, the CLI, background agents, Bugbot or service accounts. Seat licenses are shown on their own. | Cursor | request source |
| Cursor - Cost by User | Your Cursor spend, seats included, by person. Requests from service accounts and automations are shown on their own. | Cursor | user |
| Cursor - Cost per Accepted Line | Your total Cursor cost, seats included, divided by the lines of AI-written code your developers accepted. Lower means more output for the money. | Cursor | Not grouped (one Total group) |
| Cursor - Cost per Accepted Line by Billing Group | Cursor cost per accepted line of AI-written code for each billing group, so you can compare what each team gets for its spend. | Cursor | billing group |
| Cursor - Cost per Accepted Line by User | Cursor cost for each person, their seat included, per accepted line of AI-written code. Very high values point at expensive models or prompts that rarely land. | Cursor | user |
| Cursor - Cost per Agent Request | Your total Cursor cost, seats included, divided by the agent requests your developers made. Shows what each delegated task costs. | Cursor | Not grouped (one Total group) |
| Cursor - Cost per Agent Request by Billing Group | Cursor cost per agent request for each billing group, so you can compare how efficiently teams delegate work to the agent. | Cursor | billing group |
| Cursor - Cost per Agent Request by User | Cursor cost for each person, their seat included, per agent request, so you can spot where premium models make delegated work expensive. | Cursor | user |
| Cursor - Price per 1M Tokens by Model | What each model costs you per million tokens in Cursor, cached tokens included. Seat fees and included requests, which cost nothing extra, are left out. | Cursor | service |
| Cursor - Requests in Max Mode | How many Cursor requests ran in Max Mode, which uses larger context and costs more per request, against those that did not. | Cursor | Max Mode |
| Cursor - Requests by Model | How many Cursor requests each model served, so you can see which models your developers rely on. | Cursor | service |
| Cursor - Requests by Request Source | How many Cursor requests came from the editor, the CLI, background agents, Bugbot or service accounts. | Cursor | request source |
| Cursor - Requests by User | How many Cursor requests each person made. Requests from service accounts and automations are shown on their own. | Cursor | user |
| Cursor - Tokens by Billing Group | How many tokens each Cursor billing group processed, cached tokens included, so each team can see its share of AI usage. | Cursor | billing group |
| Cursor - Tokens by Model | How many tokens each model processed in Cursor, cached tokens included, so you can see where large contexts go. | Cursor | service |
| Cursor - Usage Value by Model | What your Cursor usage would cost at model prices, included usage too, by model. Compare it with your bill to see how much the seat allowance covers. | Cursor | service |
| GCP - AI Cost | Your Google Cloud AI and machine learning spend grouped by service, so you can see which AI services cost the most. | GCP | service |
| Gemini - Cache Hit Rate by Model | The share of the input tokens each Gemini model processed that came from the context cache instead of being read fresh, whether through the Gemini API or Vertex AI. A higher share means you pay the cheaper cached price for more of your input. | GCP | model |
| Gemini - Cost by Billing Item | Your Gemini spend, whether it comes through the Gemini API or through Vertex AI, by the item Google bills it on. Each item names the model and whether it is input or output tokens, so you can see which models and token types cost the most. | GCP | usage type |
| Gemini - Cost by Model | Your Gemini spend by model, whether it comes through the Gemini API or through Vertex AI, so you can see which models cost the most. Spend that is not billed per token, such as video generation, is shown on its own line. | GCP | model |
| Gemini - Cost by Project | Your Gemini spend by Google Cloud project, whether it comes through the Gemini API or through Vertex AI, so each team can see its share. | GCP | account |
| Gemini - Cost by Token Type | Your Gemini spend split into input, output and cached input tokens, whether it comes through the Gemini API or through Vertex AI, so you can see which kind of token drives the bill. Spend that is not billed per token is shown on its own line. | GCP | token type |
| Gemini - Price per 1M Tokens by Model | What each Gemini model costs you per million tokens, cached tokens included, whether through the Gemini API or Vertex AI. Spend that is not billed per token, such as video generation, is left out. | GCP | model |
| Gemini - Price per 1M Tokens by Token Type | What input, output and cached input tokens cost you per million on Gemini, whether through the Gemini API or Vertex AI, so you can see which kind of token drives the bill. | GCP | token type |
| Gemini - Tokens by Model | How many tokens each Gemini model processed, input, output and cached tokens together, whether through the Gemini API or Vertex AI, so you can see where the volume goes. | GCP | model |
| Gemini - Tokens by Token Type | How many input, output and cached input tokens Gemini processed, whether through the Gemini API or Vertex AI, so you can see how much of your volume the cache serves. | GCP | token type |
| Vertex AI - Partner Model Cost by Model | Your spend on partner models you run through Vertex AI Model Garden as a managed service, such as OpenAI gpt-oss, by model. | GCP | model |
| OpenAI - Cost by Model | Your OpenAI spend grouped by model, so you can see which models cost the most. | OpenAI | service |