Pricing model

All costs in claudex are approximate. They come from published Anthropic pricing tiers applied to the token-usage blocks recorded in each session.

Source of truth: src/types.rs, ModelPricing::for_model.

Tiers

Model tier	Input	Output	Cache write	Cache read
Opus	$15.00 / MTok	$75.00 / MTok	$18.75 / MTok	$1.50 / MTok
Sonnet	$3.00 / MTok	$15.00 / MTok	$3.75 / MTok	$0.30 / MTok
Haiku	$0.80 / MTok	$4.00 / MTok	$1.00 / MTok	$0.08 / MTok

(MTok = million tokens.)

Tier detection

The tier is chosen from a substring of the model name:

• Name contains opus → Opus.
• Name contains haiku → Haiku.
• Anything else → Sonnet (the safe fallback).

So claude-opus-4-7, claude-opus-4-6, and opus all map to Opus. claude-haiku-4-5-20251001 maps to Haiku. Unknown or missing names map to Sonnet — the middle tier — which underestimates Opus and overestimates Haiku by a small amount.

Computation

For each (session, model) row in the token_usage table:

cost = (input  × input_per_mtok
      + output × output_per_mtok
      + cache_write × cache_write_per_mtok
      + cache_read  × cache_read_per_mtok) / 1_000_000

The four token counts come from the usage block on each assistant message (Claude Code records them verbatim from the API response).

Sessions that switched models accumulate multiple rows; totals sum across them.

Why it's approximate

• No volume discounts. Priority throughput, batch pricing, etc. aren't reflected.
• No historical pricing. If tiers change, old sessions are priced at current rates. Claudex doesn't store a rate card.
• No free tier / promo credits. These are invoicing concerns; they don't show up in the API response.
• Cache-read estimate. Cache reads don't always correspond to billable tokens 1:1 in every context. Claudex prices them at the posted rate, which is a close upper bound.

For authoritative billing, use Anthropic's console. Claudex is for relative comparisons — "which project costs more", "which model tier am I leaning on", "how does this week compare to last" — where the model-agnostic math is accurate enough.

Opus:Sonnet ratio

Opus is exactly 5× Sonnet on input, output, and cache reads, and exactly 5× Sonnet on cache writes as well. If you see an Opus session that claims cost less than 5× what you'd expect from the same session run on Sonnet, check for model switching mid-session.

Rendering

• fmt_cost renders $12,345.67 with thousands separators.
• Values below one cent fall back to four decimals: $0.0042. Tiny sessions don't disappear into $0.00.
• JSON output always uses raw cost_usd floats — no formatting.