a395a25705
Bedrock multimodal embedding models (Titan and Nova) were being costed using the per-token text rate instead of the correct flat per-image rate ($0.00006/image). The pricing data was correct but never applied because image_count was never populated in prompt_tokens_details. Pass batch_data to Titan/Nova response transformers so they can count image inputs and set PromptTokensDetailsWrapper(image_count=N) on Usage, mirroring the existing Vertex AI pattern from PR #9623. Also fix the text_tokens fallback in generic_cost_per_token to not override text_tokens=0 when image_count > 0 (image-only requests). |
||
|---|---|---|
| .. | ||
| amazon_nova/chat | ||
| anthropic | ||
| azure | ||
| azure_ai | ||
| baseten/chat | ||
| bedrock | ||
| bytez/chat | ||
| chat | ||
| chatgpt | ||
| cohere | ||
| cometapi/chat | ||
| compactifai | ||
| custom_httpx | ||
| dashscope | ||
| databricks | ||
| datarobot | ||
| deepgram | ||
| deepinfra | ||
| docker_model_runner | ||
| featherless_ai/chat | ||
| fireworks_ai | ||
| gemini | ||
| github_copilot | ||
| heroku | ||
| hosted_vllm | ||
| huggingface | ||
| jina_ai/embedding | ||
| lemonade | ||
| litellm_proxy/chat | ||
| llamafile/chat | ||
| lm_studio | ||
| manus | ||
| meta_llama | ||
| minimax | ||
| mistral | ||
| moonshot | ||
| nebius | ||
| novita/chat | ||
| nscale/chat | ||
| oci/chat | ||
| ollama | ||
| openai | ||
| openai_like | ||
| openrouter | ||
| ovhcloud | ||
| perplexity | ||
| pg_vector/vector_stores | ||
| publicai | ||
| ragflow/chat | ||
| recraft | ||
| runwayml | ||
| s3_vectors | ||
| sagemaker | ||
| sambanova | ||
| sap | ||
| snowflake | ||
| stability | ||
| vercel_ai_gateway | ||
| vertex_ai | ||
| volcengine | ||
| voyage/rerank | ||
| wandb | ||
| watsonx | ||
| xai | ||
| zai | ||
| test_cache_control_and_reasoning.py | ||
| test_lifecycle_fix.py | ||
| test_oom_fixes.py | ||