Comparing
litellm_bedrock-messages-cache-prompt-double-count
(f0d2d26
) withmain
(d0e347a
)0%
- Untouched
- 16
Benchmarks
16 total
test_get_model_info_anthropic
tests/benchmarks/test_benchmarks.py+2%
85.8 µs83.8 µs
test_token_counter_long_content
tests/benchmarks/test_benchmarks.py+2%
1.7 ms1.7 ms
test_get_model_info_openai
tests/benchmarks/test_benchmarks.py+1%
89 µs88 µs
test_token_counter_multi_turn
tests/benchmarks/test_benchmarks.py0%
572.7 µs571.5 µs
test_token_counter_simple_message
tests/benchmarks/test_benchmarks.py0%
240.3 µs239.8 µs
test_get_llm_provider_anthropic
tests/benchmarks/test_benchmarks.py0%
146.8 µs146.8 µs
test_token_counter_with_tools
tests/benchmarks/test_benchmarks.py0%
418 µs418.7 µs
test_token_counter_raw_text
tests/benchmarks/test_benchmarks.py0%
187.1 µs187.6 µs
test_get_llm_provider_with_prefix
tests/benchmarks/test_benchmarks.py0%
141.8 µs142.2 µs
test_get_llm_provider_azure
tests/benchmarks/test_benchmarks.py0%
148.8 µs149.2 µs
test_get_model_cost_key_case_insensitive
tests/benchmarks/test_benchmarks.py0%
87.3 µs87.7 µs
test_cost_per_token_openai
tests/benchmarks/test_benchmarks.py0%
543.2 µs545.4 µs
test_get_model_cost_key_exact_match
tests/benchmarks/test_benchmarks.py0%
85 µs85.3 µs
test_cost_per_token_anthropic
tests/benchmarks/test_benchmarks.py0%
541.8 µs544.4 µs
test_get_llm_provider_openai
tests/benchmarks/test_benchmarks.py-1%
142.6 µs143.4 µs
test_get_model_info_with_provider
tests/benchmarks/test_benchmarks.py-1%
85.2 µs85.8 µs
Commits
Click on a commit to change the comparison rangeBase
main
d0e347a
+0.11%
fix(bedrock): avoid double-counting cache tokens in Anthropic Messages streaming usage
f0d2d26
7 days ago
by Sameerlite