bashkit
Benches

Latest benchmark snapshot

Static aggregate generated from repository result artifacts. Use the linked files for raw measurements and full eval traces.

Latest reports

Open Markdown reports

Python startup

python3: CPython (wasm) vs Monty

Open report (2026-10-07)
MeasureCPythonMonty
First callfresh process, includes runtime load18.85 ms0.575 ms
Next callsame process, new interpreter state4.412 ms0.049 ms
Warm print(1)median per call in a long-lived session4.424 ms0.018 ms
Python startup history

CPython per run

RunFirst callWarm print(1)
2026-10-07import http.client: 461.12 ms18.85 ms4.424 ms
2026-10-06import timing not recorded21.896 ms6.312 ms
Gap telemetry

Top gaps in replayed agent calls

Open report (2026-10-07)
GapCalls nowRecorded
expected 'done'parse error2338
unterminated command substitutionparse error1823
busyboxmissing command100
tar -unknown option91
unexpected end of input in [[ ]]parse error87
sudomissing command77
python3missing command66
grep Iunknown option60
gawkmissing command51
perlmissing command43
(no detail)parse error311
unexpected tokenparse error34
join ounknown option30
nodemissing command22
find -lsunknown option22
pythonmissing command21
unterminated single quoteparse error21
mawkmissing command20
gzip cunknown option20
xargs runknown option20
Gap history

Calls hitting a gap

ReplayNowRecorded
2026-10-075,206 calls, bashkit 0.18.2125319
Oils spec suite

97.3% of what real bash passes

Open report (2026-10-09)
Spec file (most misses)bashkitbash
background74.1% of 27 cases2027
bugs82.8% of 29 cases2429
builtin-trap87.9% of 33 cases2933
brace-expansion94.5% of 55 cases5255
builtin-meta83.3% of 18 cases1518
builtin-trap-bash87% of 23 cases2023
sh-options92.3% of 39 cases3639
assign-extended94.3% of 39 cases3335
builtin-read96.9% of 64 cases6264
builtin-trap-err90.9% of 23 cases2022
builtin-umask91.7% of 24 cases2224
fatal-errors60% of 5 cases35
Oils spec history

Cases passed vs real bash

RunRatebashkit / bash
2026-10-09Oils 57d3f0d088c3, bashkit 298bf8f9cfd197.3%2,618 / 2,692
2026-10-08Oils 57d3f0d088c3, bashkit 74d816bc789395.8%2,578 / 2,692
2026-10-08Oils 57d3f0d088c3, bashkit 38a65ba78b0b92.8%2,500 / 2,693
2026-10-08Oils 57d3f0d088c3, bashkit fa242de1390489%2,395 / 2,692
2026-10-08Oils 57d3f0d088c3, bashkit 26a3acc7517786%2,316 / 2,693
2026-10-08Oils 57d3f0d088c3, bashkit 017d9bfb4f3d82.1%2,211 / 2,692
2026-10-08Oils 57d3f0d088c3, bashkit 5f04b4810ec278.8%2,120 / 2,692
2026-10-08Oils 57d3f0d088c3, bashkit bae540476f2177.3%2,080 / 2,692
2026-10-08Oils 57d3f0d088c3, bashkit f86bd7c81ae374.2%1,997 / 2,693
Runtime snapshot

Latest benchmark categories

Browse benchmark runs
CategoryCasesLast run
startupSmall commands where interpreter startup dominates runtime.40.016 msbash median: 4.769 ms
arithmeticInteger math, substitutions, and expression-heavy shell snippets.60.018 msbash median: 4.313 ms
stringsString expansion, pattern handling, and text manipulation.80.018 msbash median: 4.378 ms
variablesVariable assignment, lookup, expansion, and environment handling.80.019 msbash median: 4.096 ms
arraysIndexed array reads, writes, expansion, and iteration.60.019 msbash median: 4.064 ms
pipesPipeline construction, streaming, and command chaining.60.021 msbash median: 5.898 ms
subshellCommand substitution and nested shell execution paths.60.024 msbash median: 5.484 ms
controlConditionals, loops, case statements, and branching scripts.90.025 msbash median: 3.812 ms
Eval pressure

Lowest eval categories

Browse eval runs
Latest LLM eval93%54/58 tasks
CategoryPassedPass rate
system_info1/250%tasks passed
file_operations3/466.7%tasks passed
scripting5/768.6%tasks passed
json_processing8/8100%tasks passed
data_transformation6/6100%tasks passed
complex_tasks6/6100%tasks passed
text_processing6/6100%tasks passed
pipelines5/5100%tasks passed