Squeezy vs external baseline (Mini)
Both on gpt-5.4-mini, median of 10 runs. Every language came in cheaper at equal-or-better recall.
benchmarks
The saving, measured.
Two boards pit Squeezy against an external coding-agent baseline on the same code-understanding tasks: same prompt, same model, same grader. Squeezy answers from its local knowledge of the code, and at the median it spent 60 cents for every dollar the baseline spent.
Squeezy vs external baseline (Haiku)
Both on claude-haiku-4-5, median of 10 runs. Every language came in cheaper at equal-or-better recall.
language ratios
Cost ratio by language.
Ratio is Squeezy's cost divided by the baseline's: 0.60 means Squeezy spent 60 cents where the baseline spent a dollar. Cost only counts when the answer passes the board's recall rule, so a cheap wrong answer can't win. The 15 rows are the general-purpose programming languages in this suite; markup and documentation formats such as CSS, HTML, and Markdown are not scored. C and C++, and JavaScript and TypeScript, count as separate rows.
Mini tier
15 language tasks
Dart 68% lower
0.32 $0.0161
JS 68% lower
0.32 $0.0080
Java 66% lower
0.34 $0.0373
TS 66% lower
0.34 $0.0135
Go 58% lower
0.42 $0.0146
Ruby 52% lower
0.48 $0.0100
C 43% lower
0.57 $0.0207
Python 40% lower
0.60 $0.0079
C++ 30% lower
0.70 $0.0576
Swift 28% lower
0.72 $0.0123
PHP 20% lower
0.80 $0.0317
C# 18% lower
0.82 $0.0291
Kotlin 18% lower
0.82 $0.0353
Rust 15% lower · recall 93.3%
0.85 $0.0246
Scala 14% lower
0.86 $0.0355
Haiku tier
15 language tasks
Go 86% lower
0.14 $0.0175
Python 72% lower
0.28 $0.0092
Scala 71% lower
0.29 $0.0803
Dart 65% lower
0.35 $0.0533
C# 62% lower
0.38 $0.0548
Rust 60% lower · recall 90%
0.40 $0.0576
JS 58% lower
0.42 $0.0202
Kotlin 57% lower
0.43 $0.0657
Ruby 57% lower
0.43 $0.0330
C++ 55% lower
0.45 $0.1002
C 49% lower
0.51 $0.0549
TS 44% lower
0.56 $0.1051
Java 39% lower
0.61 $0.1382
PHP 26% lower
0.74 $0.0961
Swift 22% lower
0.78 $0.0228
This is a fixed publication snapshot built from medians of 10 runs. Each baseline uses the same repository question, grader, model tier, and pricing assumptions.
Prompt and repo are matched
Each row compares answers to the same real-world code question, scored by the same grader.
Multiple runs reduce noise
Every figure is the median of 10 runs, because single agent runs vary in both cost and recall.
What the numbers cover
These results come from this task suite on these model tiers. Your own bill depends on the models you pick, your provider's pricing, and the shape of your work.