Skip to content

fix(google-ai): bill Gemini cached prompt tokens separately #1496

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
pcbeingused333 wants to merge 1 commit into
base: master
Choose a base branch
from
Open

fix(google-ai): bill Gemini cached prompt tokens separately #1496

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 6 additions & 3 deletions src/Plugins/BotSharp.Plugin.GoogleAI/Providers/Chat/ChatCompletionProvider.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -126,7 +126,8 @@ public async Task<RoleDialogModel> GetChatCompletions(Agent agent, List<RoleDial
Prompt = prompt,
Provider = Provider,
Model = _model,
TextInputTokens = response?.UsageMetadata?.PromptTokenCount ?? 0,
TextInputTokens = ProviderHelper.GetInputTokens(response?.UsageMetadata),
CachedTextInputTokens = ProviderHelper.GetCachedInputTokens(response?.UsageMetadata),
TextOutputTokens = ProviderHelper.GetOutputTokens(response?.UsageMetadata)
});
}
Expand Down Expand Up @@ -183,7 +184,8 @@ public async Task<bool> GetChatCompletionsAsync(Agent agent, List<RoleDialogMode
Prompt = prompt,
Provider = Provider,
Model = _model,
TextInputTokens = response?.UsageMetadata?.PromptTokenCount ?? 0,
TextInputTokens = ProviderHelper.GetInputTokens(response?.UsageMetadata),
CachedTextInputTokens = ProviderHelper.GetCachedInputTokens(response?.UsageMetadata),
TextOutputTokens = ProviderHelper.GetOutputTokens(response?.UsageMetadata)
});
}
Expand Down Expand Up @@ -446,7 +448,8 @@ public async Task<RoleDialogModel> GetChatCompletionsStreamingAsync(Agent agent,
Prompt = prompt,
Provider = Provider,
Model = _model,
TextInputTokens = tokenUsage?.PromptTokenCount ?? 0,
TextInputTokens = ProviderHelper.GetInputTokens(tokenUsage),
CachedTextInputTokens = ProviderHelper.GetCachedInputTokens(tokenUsage),
TextOutputTokens = ProviderHelper.GetOutputTokens(tokenUsage)
});
}
Expand Down
14 changes: 14 additions & 0 deletions src/Plugins/BotSharp.Plugin.GoogleAI/Providers/ProviderHelper.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,4 +26,18 @@ public static long GetOutputTokens(GenerativeAI.Types.UsageMetadata? usage)
{
return (usage?.CandidatesTokenCount ?? 0) + (usage?.ThoughtsTokenCount ?? 0);
}

/// <summary>
/// Gemini's prompt count includes the tokens served from the context cache,
/// which are billed at a separate rate, so they are subtracted here.
/// </summary>
public static long GetInputTokens(GenerativeAI.Types.UsageMetadata? usage)
{
return Math.Max((usage?.PromptTokenCount ?? 0) - GetCachedInputTokens(usage), 0);
}

public static long GetCachedInputTokens(GenerativeAI.Types.UsageMetadata? usage)
{
return usage?.CachedContentTokenCount ?? 0;
}
}
3 changes: 2 additions & 1 deletion src/Plugins/BotSharp.Plugin.GoogleAI/Providers/Text/TextCompletionProvider.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,7 +62,8 @@ public async Task<string> GetCompletion(string text, string agentId, string mess
Prompt = text,
Provider = Provider,
Model = _model,
TextInputTokens = response?.UsageMetadata?.PromptTokenCount ?? 0,
TextInputTokens = ProviderHelper.GetInputTokens(response?.UsageMetadata),
CachedTextInputTokens = ProviderHelper.GetCachedInputTokens(response?.UsageMetadata),
TextOutputTokens = ProviderHelper.GetOutputTokens(response?.UsageMetadata)
});
}
Expand Down
18 changes: 18 additions & 0 deletions tests/BotSharp.Core.UnitTests/Plugins/GoogleAiUsageTests.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,4 +20,22 @@ public void Output_tokens_without_thinking_are_the_candidates_count()
Assert.Equal(20, ProviderHelper.GetOutputTokens(new UsageMetadata { CandidatesTokenCount = 20 }));
Assert.Equal(0, ProviderHelper.GetOutputTokens(null));
}

[Fact]
public void Cached_tokens_are_not_counted_as_fresh_input()
{
var usage = new UsageMetadata { PromptTokenCount = 1000, CachedContentTokenCount = 800 };

Assert.Equal(200, ProviderHelper.GetInputTokens(usage));
Assert.Equal(800, ProviderHelper.GetCachedInputTokens(usage));
}

[Fact]
public void Input_tokens_without_cache_are_the_prompt_count()
{
Assert.Equal(50, ProviderHelper.GetInputTokens(new UsageMetadata { PromptTokenCount = 50 }));
Assert.Equal(0, ProviderHelper.GetCachedInputTokens(new UsageMetadata { PromptTokenCount = 50 }));
Assert.Equal(0, ProviderHelper.GetInputTokens(null));
Assert.Equal(0, ProviderHelper.GetCachedInputTokens(null));
}
}
Loading