Spaces:
Paused
Paused
File size: 2,477 Bytes
3ed7ad0 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 | """Price table and cost derivation.
Claude rows are Anthropic first-party list prices (cached model table, 2026-10-06): input and
output as listed, cache read as listed (0.1x input unless noted), 5-minute cache write 1.25x
input. Haiku 5.5 matches Claude Code's own `total_cost_usd` (tests/test_prices.py). The Codex row
is still a placeholder. Override any row with a JSON file whose path is in FLOCK_PRICES:
{"claude-opus-5-5": {"input": 15, "output": 75, "cache_read": 1.5, "cache_write": 18.75}}
All prices are USD per 1M tokens. Reasoning tokens are assumed to be included in the
output count the harness reports, so they are not billed separately.
"""
from __future__ import annotations
import json
import os
from dataclasses import dataclass
from functools import lru_cache
from flock.models import TokenUsage
@dataclass(frozen=True)
class Price:
input: float
output: float
cache_read: float
cache_write: float
DEFAULT_MODEL = "default"
# See module docstring.
PRICES: dict[str, Price] = {
"claude-opus-5-5": Price(4.0, 20.0, 0.2, 5.0), # cache read listed at $0.20
"claude-sonnet-5-5": Price(2.0, 10.0, 0.2, 2.5),
"claude-haiku-5-5": Price(0.10, 0.50, 0.01, 0.125),
"gpt-5.5-codex": Price(1.25, 10.0, 0.125, 0.0),
DEFAULT_MODEL: Price(15.0, 75.0, 1.5, 18.75), # unknown models priced high, not free
}
@lru_cache(maxsize=1)
def price_table() -> dict[str, Price]:
table = dict(PRICES)
path = os.environ.get("FLOCK_PRICES")
if path:
with open(path, encoding="utf-8") as f:
for model, p in json.load(f).items():
table[model] = Price(**p)
return table
def price_for(model: str, table: dict[str, Price] | None = None) -> Price:
"""Exact match, else the longest table key that prefixes the model id, else default."""
table = table or price_table()
if model in table:
return table[model]
prefixes = [k for k in table if k != DEFAULT_MODEL and model.startswith(k)]
return table[max(prefixes, key=len)] if prefixes else table[DEFAULT_MODEL]
def cost_usd(tokens: dict[str, TokenUsage], table: dict[str, Price] | None = None) -> float:
total = 0.0
for model, u in tokens.items():
p = price_for(model, table)
total += (
u.input * p.input
+ u.output * p.output
+ u.cache_read * p.cache_read
+ u.cache_write * p.cache_write
) / 1_000_000
return round(total, 6)
|