Add API-equivalent pricing to llm-cost CLI#8
Merged
Conversation
Ports the rate table that lived in software-factory's cost-attribution
bounded context into the OSS package, plus the calculation logic to
turn token-bucket totals into USD per provider per-issue.
New library exports:
- PRICING_TABLE the rate data (Anthropic + OpenAI)
- calculateCost(model, buckets) per-bucket cost breakdown + total
- ratesForModel(model) lookup with normalization
- normalizeModelName(model) strip `-YYYYMMDD`, `-latest`, dash-to-dot
- hypotheticalNoteFor(prov, plan) UI text helper
- daysSincePricingVerified() freshness check
- isPricingStale() true after STALE_AFTER_DAYS (90)
- STALE_AFTER_DAYS
CLI changes:
- Default output now includes an "API-equivalent pricing" block
per provider, with per-bucket math:
input uncached $7.59 (1.5M × $5.00/1M)
cache read $25.51 (51.0M × $0.500/1M)
output (visible) $1.34 (44.7K × $30.00/1M)
output (reasoning) $0.56 (18.6K × $30.00/1M)
total API cost $35.00 [hypothetical — your Codex Pro plan covers this]
The "[hypothetical — ...]" suffix is plan-aware (uses the planType
from the Codex quota sample when present).
- The Quota line now reports the per-issue delta:
5h primary 58% → 64% used (peak 64%, this issue moved +6.0 pp)
- --no-pricing flag suppresses the pricing block.
- --json output gains a `pricing` field per provider when rates known.
- Warns when the rate table is more than 90 days old.
Honest framing in the README: this is a *counterfactual* (what the
token volume would have cost on pay-as-you-go API), not actual spend
on a subscription plan. The Codex quota readout is closer to your
real marginal cost; the dollar number is for cross-plan comparison.
Tested against real data — the EPAC-1940 Codex run on this machine
prices out to $35.00 (1.5M uncached + 51M cached + 44.7K visible +
18.6K reasoning), which matches the hand-computed value. A test in
pricing.test.mjs pins that exact total to catch any future regression.
51 of 51 package tests pass (was 33; added 18 pricing tests).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The OSS `llm-cost` CLI now shows dollar-equivalent cost per bucket alongside the raw token counts. Rates come from a built-in table sourced from anthropic.com/pricing and platform.openai.com/docs/pricing, verified 2026-05-22.
What `./cost EPAC-1940` shows now
```
CODEX (1 session)
Models: gpt-5.5
Turns: 341
First → last: 2026-05-18T19:56:54.104Z → 2026-05-18T21:50:34.098Z
Wall clock span: 1h 53m 40s
Tokens:
input uncached 1,517,206
cache read 51,024,768
output (visible) 44,683
output (reasoning) 18,649
─────────────────────────────────
grand total 52,605,306
API-equivalent pricing (gpt-5.5 @ rates verified 2026-05-22):
input uncached $7.59 (1.5M × $5.00/1M)
cache read $25.51 (51.0M × $0.500/1M)
output (visible) $1.34 (44.7K × $30.00/1M)
output (reasoning) $0.56 (18.6K × $30.00/1M)
───────────────────────────────────────────
total API cost $35.00 [hypothetical — your Codex Pro plan covers this]
Quota (plan_type=pro, 341 samples):
5h primary 58% → 64% used (peak 64%, this issue moved +6.0 pp)
7d secondary 56% → 57% used (peak 57%, this issue moved +1.0 pp)
```
EPAC-1940 prices out to $35.00 hypothetical API equivalent but actually moved the user's Codex Pro 7-day quota by +1.0 pp (~1/100th of the weekly allowance). Both readouts in one shot.
Honest framing
The README and the in-CLI `[hypothetical — ...]` suffix make clear that the dollar number is a counterfactual — what the token volume would have cost on pay-as-you-go API, not actual spend on a subscription plan. The Codex quota readout is closer to real marginal cost on Codex Pro / Codex Business; the dollar total is for cross-plan comparison and external audiences who think in dollars.
Implementation
New files:
CLI additions:
Test plan
Follow-ups (not in this PR)