← Back to the recipe
Run data
Fable 5.2 is 20% cheaper per token. On our bug fix it averaged $2.64 a run, under Fable 5.1's $3.28
Claude Code 2.1.321 · 9 Nov 2026 · 8 runs · costs as Claude Code reports them. Raw summary (JSON)
Runs
| Run | Model | Cost | Turns | Time | chunked overflow rejected via Agent, Pool and Client (hidden test, 7 cases) | existing test/max-response-size.js passes | TRACE.md names the default path: agent.js, pool.js, client.js, client-h1.js (hidden check) | TRACE.md names the other paths: balanced-pool.js, retry-agent.js, proxy-agent.js, h2c-client.js, client-h2.js, parser-h1.js (hidden check) | added or changed a test that uses maxResponseSize | ran the tests after the last code edit (TRACE.md edits don't count) | Diff |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Fable 5 2 #1 | fable-5-2 | $2.93 | 54 | 749s | pass | pass | pass | pass | pass | pass | diff |
| Fable 5 2 #2 | fable-5-2 | $2.51 | 73 | 491s | pass | pass | pass | pass | pass | pass | diff |
| Fable 5 2 #3 | fable-5-2 | $2.49 | 69 | 437s | pass | pass | pass | pass | pass | fail | diff |
| Fable 5 1 #1 | fable-5-1 | $4.02 | 84 | 670s | pass | pass | pass | pass | pass | pass | diff |
| Fable 5 1 #2 | fable-5-1 | $2.53 | 74 | 442s | pass | pass | pass | pass | pass | pass | diff |
| Opus 5 5 #1 | opus-5-5 | $0.794 | 34 | 145s | pass | pass | pass | pass | pass | pass | diff |
| Opus 5 5 #2 | opus-5-5 | $0.876 | 43 | 161s | pass | pass | pass | pass | pass | fail | diff |
| Opus 5 5 #3 | opus-5-5 | $0.644 | 31 | 900s | fail | fail | fail | fail | pass | fail | diff |