github-featured / comet-ml--opik

发布

0 发布 8,340 Git 标签

RSS 订阅源
  • 1.10.43-4534 25460fc90a
    比较

    [issue-5618] [BE] fix: bill LiteLLM OTel cache tokens at cache-read rate (#5643)

    frostbyte_neo 发布于 2026-03-17 15:15:25 +00:00

    • [issue-5618] [BE] fix: add OTel cache token key fallback in SpanCostCalculator

    LiteLLM via OpenTelemetry emits cache token counts as bare keys (e.g.
    cache_read_input_tokens) without the original_usage. prefix used by
    the Python SDK. Add fallbacks to these OTel bare keys in both
    textGenerationWithCacheCostOpenAI and
    textGenerationWithCachedTokensNotIncludedInCost (used by Anthropic and
    Bedrock) so cache-read tokens are billed at the cache-read rate rather
    than the full input rate.

    Co-Authored-By: Claude Sonnet 4.6 noreply@anthropic.com

    • [issue-3447] [BE] Polish SpanCostCalculatorTest per review feedback
    • Merge two OpenAI cache cost @Tests into @ParameterizedTest provideOpenAICacheCostCases
    • Merge Anthropic and Bedrock OTel fallback @Tests into @ParameterizedTest provideAnthropicBedrockCacheCostCases
    • Replace BigDecimal arithmetic expected values with direct strings, keeping math in comments

    Co-Authored-By: Claude Sonnet 4.6 noreply@anthropic.com

    Co-authored-by: Douglas Blank doug@comet.com
    Co-authored-by: Claude Sonnet 4.6 noreply@anthropic.com

    下载附件