From 92b30bcad58e5c15f85d64d5a829c8d2a16a3d95 Mon Sep 17 00:00:00 2001 From: Alex Schapiro Date: Wed, 28 Oct 2026 18:03:04 +0000 Subject: [PATCH] fix(STR-1654): price LLM usage from model config --- docs/advanced/configuration.mdx | 12 ++++ strix/report/usage.py | 69 ++++++++++++++++++- tests/test_hooks.py | 32 +++++++++ tests/test_pricing.py | 116 ++++++++++++++++++++++++++++++++ 4 files changed, 227 insertions(+), 2 deletions(-) diff --git a/docs/advanced/configuration.mdx b/docs/advanced/configuration.mdx index dfd28d23f..76d434ebc 100644 --- a/docs/advanced/configuration.mdx +++ b/docs/advanced/configuration.mdx @@ -19,6 +19,18 @@ Configure Strix using environment variables or a config file. Custom API base URL. Also accepts `OPENAI_API_BASE`, `LITELLM_BASE_URL`, or `OLLAMA_API_BASE`. + + Price per 1M input tokens in USD. When set, Strix computes cost and `--max-budget` from these prices instead of the provider-reported or LiteLLM cost. Use it for gateway or self-hosted model ids that LiteLLM can't price. + + + + Price per 1M output tokens in USD. When set, Strix computes cost and `--max-budget` from these prices instead of the provider-reported or LiteLLM cost. Use it for gateway or self-hosted model ids that LiteLLM can't price. + + + + Price per 1M cached input tokens in USD. When set, Strix computes cost and `--max-budget` from these prices instead of the provider-reported or LiteLLM cost. If unset, cached tokens use the input-token price. + + Select the OpenAI API path for the model: `responses` or `chat_completions`. By default the model decides: Responses when LiteLLM's model catalog lists diff --git a/strix/report/usage.py b/strix/report/usage.py index bb20b254d..4e03944b1 100644 --- a/strix/report/usage.py +++ b/strix/report/usage.py @@ -3,6 +3,8 @@ from __future__ import annotations import logging +import math +import os from typing import Any from agents.usage import Usage, deserialize_usage, serialize_usage @@ -13,6 +15,53 @@ from strix.report.pricing import resolve_litellm_model logger = logging.getLogger(__name__) +_PRICE_ENV = ( + "STRIX_LLM_INPUT_COST_PER_MTOK", + "STRIX_LLM_OUTPUT_COST_PER_MTOK", + "STRIX_LLM_CACHED_INPUT_COST_PER_MTOK", +) +_warned_invalid_prices: set[str] = set() + + +def _configured_prices() -> tuple[float | None, float | None, float | None]: + prices: list[float | None] = [] + for name in _PRICE_ENV: + raw = os.environ.get(name) + if raw is None or raw == "": + prices.append(None) + continue + try: + price = float(raw) + except ValueError: + price = math.nan + if not math.isfinite(price) or price < 0: + if name not in _warned_invalid_prices: + logger.warning("Ignoring invalid %s; expected a finite non-negative rate", name) + _warned_invalid_prices.add(name) + prices.append(None) + continue + prices.append(price) + return prices[0], prices[1], prices[2] + + +def _configured_usage_cost( + usage: Usage, + prices: tuple[float | None, float | None, float | None], +) -> float: + input_rate, output_rate, cached_rate = prices + input_tokens = max(0, _int_or_zero(getattr(usage, "input_tokens", 0))) + cached_details = getattr(usage, "input_tokens_details", None) + cached_tokens = max(0, _int_or_zero(getattr(cached_details, "cached_tokens", 0))) + cached_tokens = min(cached_tokens, input_tokens) + output_tokens = max(0, _int_or_zero(getattr(usage, "output_tokens", 0))) + billable_input = input_tokens - cached_tokens + cost = ( + billable_input * (input_rate or 0) + + cached_tokens * (cached_rate if cached_rate is not None else input_rate or 0) + + output_tokens * (output_rate or 0) + ) / 1_000_000 + return _round_cost(cost) + class ProviderUsage(BaseModel): """Running totals for one upstream provider OpenRouter routed calls to.""" @@ -114,11 +163,23 @@ class LLMUsageLedger: def total_cost(self) -> float: if self.zero_cost: return 0.0 + prices = _configured_prices() + if any(price is not None for price in prices): + return _configured_usage_cost(self._total_usage, prices) return _round_cost(self._observed_cost if self._has_observed_cost else self._estimated_cost) def to_record(self) -> dict[str, Any]: record = serialize_usage(self._total_usage) - record["cost"] = self.total_cost + prices = _configured_prices() + configured = any(price is not None for price in prices) + total_cost = ( + 0.0 + if self.zero_cost + else _configured_usage_cost(self._total_usage, prices) + if configured + else self.total_cost + ) + record["cost"] = total_cost record["providers"] = {name: tally.model_dump() for name, tally in self._providers.items()} record["agents"] = [] @@ -128,7 +189,11 @@ class LLMUsageLedger: usage = self._agent_usage[agent_id] metadata = self._agent_metadata.get(agent_id, {}) agent_cost = ( - self.total_cost * (agent_tokens[agent_id] / total_tokens) if total_tokens else 0.0 + _configured_usage_cost(usage, prices) + if configured and not self.zero_cost + else total_cost * (agent_tokens[agent_id] / total_tokens) + if total_tokens + else 0.0 ) agent_record = serialize_usage(usage) diff --git a/tests/test_hooks.py b/tests/test_hooks.py index d19c4b250..9fca882e4 100644 --- a/tests/test_hooks.py +++ b/tests/test_hooks.py @@ -2,10 +2,12 @@ from __future__ import annotations +from types import SimpleNamespace from typing import Any from unittest.mock import MagicMock, patch import pytest +from agents.usage import Usage from strix.core.agents import AgentCoordinator from strix.core.hooks import ( @@ -15,6 +17,7 @@ from strix.core.hooks import ( SubagentBudgetReservedError, recomputed_budget_flags, ) +from strix.report.usage import LLMUsageLedger def _make_hooks(max_budget: float | None) -> ReportUsageHooks: @@ -85,6 +88,35 @@ async def test_over_budget_raises() -> None: await hooks.on_llm_end(_make_context(), MagicMock(), MagicMock()) +@pytest.mark.asyncio +async def test_manual_rates_make_unpriced_model_trip_budget( + monkeypatch: pytest.MonkeyPatch, +) -> None: + monkeypatch.setenv("STRIX_LLM_INPUT_COST_PER_MTOK", "10") + monkeypatch.delenv("STRIX_LLM_OUTPUT_COST_PER_MTOK", raising=False) + monkeypatch.delenv("STRIX_LLM_CACHED_INPUT_COST_PER_MTOK", raising=False) + usage = Usage() + usage.requests = 1 + usage.input_tokens = 1000 + usage.total_tokens = 1000 + ledger = LLMUsageLedger() + + class ReportState: + def record_sdk_usage(self, **kwargs: Any) -> None: + ledger.record(**kwargs) + + def get_total_llm_cost(self) -> float: + return ledger.total_cost + + hooks = _make_hooks(0.005) + response: Any = SimpleNamespace(usage=usage) + with ( + patch("strix.core.hooks.get_global_report_state", return_value=ReportState()), + pytest.raises(BudgetExceededError), + ): + await hooks.on_llm_end(_make_context(), MagicMock(), response) + + @pytest.mark.asyncio async def test_budget_check_uses_live_cost_accessor() -> None: # The check must read the live ledger, not the persisted run-record snapshot, diff --git a/tests/test_pricing.py b/tests/test_pricing.py index 8916982d3..f27d8a9a7 100644 --- a/tests/test_pricing.py +++ b/tests/test_pricing.py @@ -1,5 +1,6 @@ from __future__ import annotations +from typing import TYPE_CHECKING from unittest.mock import patch import litellm @@ -9,6 +10,10 @@ from strix.report.pricing import resolve_litellm_model from strix.report.usage import LLMUsageLedger +if TYPE_CHECKING: + import pytest + + def test_resolves_common_bare_model_names() -> None: resolve_litellm_model.cache_clear() assert resolve_litellm_model("deepseek-v4-flash") == "deepseek/deepseek-v4-flash" @@ -60,6 +65,117 @@ def test_ledger_prefers_observed_cost_over_estimate() -> None: assert ledger.total_cost == 0.17 +def test_unpriced_model_without_manual_rates_costs_zero(monkeypatch: pytest.MonkeyPatch) -> None: + for name in ( + "STRIX_LLM_INPUT_COST_PER_MTOK", + "STRIX_LLM_OUTPUT_COST_PER_MTOK", + "STRIX_LLM_CACHED_INPUT_COST_PER_MTOK", + ): + monkeypatch.delenv(name, raising=False) + usage = Usage() + usage.requests = 1 + usage.input_tokens = 1000 + usage.output_tokens = 200 + usage.total_tokens = 1200 + ledger = LLMUsageLedger() + + ledger.record(agent_id="a", usage=usage, model="gateway/unpriced-model") + + assert ledger.total_cost == 0.0 + + +def test_manual_rates_price_usage_and_override_observed_cost( + monkeypatch: pytest.MonkeyPatch, +) -> None: + monkeypatch.setenv("STRIX_LLM_INPUT_COST_PER_MTOK", "2") + monkeypatch.setenv("STRIX_LLM_OUTPUT_COST_PER_MTOK", "8") + monkeypatch.delenv("STRIX_LLM_CACHED_INPUT_COST_PER_MTOK", raising=False) + usage = Usage() + usage.requests = 1 + usage.input_tokens = 1000 + usage.output_tokens = 200 + usage.total_tokens = 1200 + ledger = LLMUsageLedger() + + ledger.record(agent_id="a", usage=usage, model="gateway/unpriced-model") + ledger.record_observed_cost(0.17) + + assert ledger.total_cost == 0.0036 + assert ledger.to_record()["agents"][0]["cost"] == 0.0036 + + +def test_cached_tokens_fall_back_to_input_rate(monkeypatch: pytest.MonkeyPatch) -> None: + monkeypatch.setenv("STRIX_LLM_INPUT_COST_PER_MTOK", "5") + monkeypatch.setenv("STRIX_LLM_OUTPUT_COST_PER_MTOK", "7") + monkeypatch.delenv("STRIX_LLM_CACHED_INPUT_COST_PER_MTOK", raising=False) + usage = Usage() + usage.requests = 1 + usage.input_tokens = 10 + usage.output_tokens = 1 + usage.total_tokens = 11 + usage.input_tokens_details.cached_tokens = 2 + ledger = LLMUsageLedger() + + ledger.record(agent_id="a", usage=usage, model="gateway/unpriced-model") + + assert ledger.total_cost == 0.000057 + + +def test_hydrated_usage_uses_configured_rates_not_persisted_cost( + monkeypatch: pytest.MonkeyPatch, +) -> None: + monkeypatch.setenv("STRIX_LLM_INPUT_COST_PER_MTOK", "2") + monkeypatch.setenv("STRIX_LLM_OUTPUT_COST_PER_MTOK", "8") + usage = Usage() + usage.requests = 1 + usage.input_tokens = 1000 + usage.output_tokens = 200 + usage.total_tokens = 1200 + original = LLMUsageLedger() + original.record(agent_id="a", usage=usage, model="gateway/unpriced-model") + persisted = original.to_record() + persisted["cost"] = 42.0 + hydrated = LLMUsageLedger() + + hydrated.hydrate(persisted) + + assert hydrated.total_cost == 0.0036 + + +def test_invalid_manual_rate_is_ignored( + monkeypatch: pytest.MonkeyPatch, + caplog: pytest.LogCaptureFixture, +) -> None: + monkeypatch.setenv("STRIX_LLM_INPUT_COST_PER_MTOK", "invalid") + monkeypatch.setenv("STRIX_LLM_OUTPUT_COST_PER_MTOK", "10") + monkeypatch.delenv("STRIX_LLM_CACHED_INPUT_COST_PER_MTOK", raising=False) + usage = Usage() + usage.requests = 1 + usage.input_tokens = 1000 + usage.output_tokens = 200 + usage.total_tokens = 1200 + ledger = LLMUsageLedger() + + ledger.record(agent_id="a", usage=usage, model="gateway/unpriced-model") + + assert ledger.total_cost == 0.002 + assert "Ignoring invalid STRIX_LLM_INPUT_COST_PER_MTOK" in caplog.text + + +def test_zero_cost_overrides_manual_rates(monkeypatch: pytest.MonkeyPatch) -> None: + monkeypatch.setenv("STRIX_LLM_INPUT_COST_PER_MTOK", "10") + usage = Usage() + usage.requests = 1 + usage.input_tokens = 1000 + usage.total_tokens = 1000 + ledger = LLMUsageLedger() + ledger.zero_cost = True + + ledger.record(agent_id="a", usage=usage, model="gateway/unpriced-model") + + assert ledger.total_cost == 0.0 + + def test_hydrated_estimate_continues_accumulating_new_estimates() -> None: usage = Usage() usage.requests = 1