Cloud baseline

Haiku 4.5 — 0.4615 on Gauntlet 5.3

us.anthropic.claude-haiku-4-5-20251001-v1:0 · scorer sb-5.3

Claude Haiku 4.5Aug 18, 2026
Overall
0.4615
Hard
0.17
Excellent
no
Benchmark
Gauntlet 5.3

Token usage

Tokens: at least 5.6M in · 28.5k out (lower bound, harness session counters)

5,634,367 tokens over 72 model calls: every call of the run, the model's own and the harness's helper calls, counted from harness session counters. These are a lower bound.

The harness's session counters for this run (the usage AWS Bedrock returned per response, summed); calls counted from the stored transcript (approximate). Prompt caching was off on Bedrock for Claude then, so no cached tokens. Bedrock keeps no per-call bill to query. Matched to build r1 by its score (the median rep); medium confidence in that match.

  • us.anthropic.claude-haiku-4-5-20251001-v1:0

    In 5.6M · Out 28.5k · 72 calls

Tier breakdown

A · structure & runtime0.83
B · behaviour0.36
C · vendor contract0.10
D · finesse0.63

Scoring detail

How this score was built

Per-check scorer output, as posted by the app. Expand a section to inspect its checks.

ComponentWeightEarnedPoints
Core (tiers A–D)60%0.4680.281
A · Structure25% of core0.830.125
B · Behaviour30% of core0.360.066
C · Vendor contract25% of core0.100.015
D · Finesse20% of core0.630.075
Journey (J)15%0.6500.098
Visual (V)10%0.3330.033
Performance (P)5%0.6670.033
Hard blocks10%0.1670.017
Total0.4615

Run notes and corrections

Anthropic baseline — the frozen v2 spec build, rescored by the sb-5.3 scorer. Screenshots are the scorer's own render-gate captures of the built app.

Screenshots

Recorded views of the built application. Captions identify the available captures; the number of images does not imply a number of repair rounds.

1First render — before repairs
2Final render
3After sync
4Mobile · 375px

Run details

Model
us.anthropic.claude-haiku-4-5-20251001-v1:0

How the benchmark is scored