mirror of
https://github.com/BerriAI/litellm.git
synced 2026-11-20 10:02:54 +00:00
Adds 36 apimodels/<model> chat rows to model_prices_and_context_window.json and the backup copy, so spend tracking prices apimodels calls instead of logging them at zero Rates are the per token prices apimodels.app bills, as listed on https://apimodels.app/pricing and on each model page linked in the row's source field. Context limits and the function calling, vision, reasoning and prompt caching flags are copied from the model vendor's own row The rows carry the same long prompt and off-peak rules the gateway bills with. The gpt-6 and gpt-5.6 rows rebill the whole request at 2x input and cache read and 1.5x output above 272k prompt tokens, claude-haiku-5-5 bills every rate at 5x above 100k, qwen3.7-plus bills 3x above 262,144 tokens through tiered_pricing, and the two DeepSeek Flash rows bill half rate outside 01:00-04:00 and 06:00-10:00 UTC on weekdays Two cost tests check that an apimodels row bills cached tokens at its cache read rate and that a prompt over 272k tokens moves the whole request to the long context rates |
||
|---|---|---|
| .. | ||
| _support | ||
| base_sdk_tests | ||
| benchmarks | ||
| code_coverage_tests | ||
| documentation_tests | ||
| e2e | ||
| e2e_harness | ||
| fixtures | ||
| harness_e2e | ||
| integration | ||
| rust-python-harness | ||
| test_litellm_rust | ||
| typing | ||
| unit | ||
| windows_tests | ||
| __init__.py | ||
| _fake_openai_endpoint_server.py | ||
| _flush_vcr_cache.py | ||
| _live_test_helpers.py | ||
| _master_key.py | ||
| _openai_record_replay_proxy.py | ||
| _process_helpers.py | ||
| _vcr_conftest_common.py | ||
| _vcr_redis_persister.py | ||
| _wait_helpers.py | ||
| _ws_vcr.py | ||
| AGENTS.md | ||
| capturing_transport.py | ||
| eval_swe_bench.py | ||
| fake_openai_endpoint.py | ||
| gettysburg.wav | ||
| large_text.py | ||
| openai_batch_completions.jsonl | ||
| pyrightconfig.json | ||
| README.MD | ||
| test_logging.conf | ||
| test_rust_python_harness.py | ||
| VCR_TESTING.md | ||
| white_100x100.png | ||
In total litellm runs 1000+ tests
[02/20/2025] Update:
To make it easier to contribute and map what behavior is tested,
we've started mapping the litellm directory in tests/unit
This folder can only run mock tests.