rhysd/kiro-editor | ProgramBench

A small terminal UTF-8 text editor written in Rust 📝🦀

589

Generated Behavioral Tests

99.8%

Best Score

Claude Opus 5.5 (xhigh)

Results by Model

Hover for details · Click a model to open it · Drag to zoom the x-axis · Shift-drag to pan · The line marks the Pareto frontier (best score per cost)

30 runs

#	Model	Score	Cost	Calls
1	Claude Opus 5.5 (xhigh)	99.8%	$65.16	481	trace →
2	Muse Wonder Full	99.8%	TBA	421	trace →
3	GPT 6 Astra (max)	99.3%	$14.13	69	trace →
4	Muse Wonder (xhigh)	99.2%	$48.06	391	trace →
5	Claude Opus 5 (xhigh)	99.2%	$101.86	375	trace →
6	Claude Fable 5.1 (xhigh)	98.8%	$80.39	417	trace →
7	GPT-5.6 Sol (xhigh)	96.8%	$9.58	65	trace →
8	GPT 5.5 (xhigh)	95.2%	$9.89	82	trace →
9	Muse Spark 1.3 (max)	95.1%	$5.77	303	trace →
10	Claude Opus 4.8 (xhigh)	93.4%	$49.47	217	trace →
11	Gemini 3.7 Flash	93.4%	$3.25	134	trace →
12	Claude Opus 4.6	93.2%	$15.07	329	trace →
13	GLM-5.2	92.9%	$33.24	257	trace →
14	Muse Spark 1.2 (xhigh)	91.7%	$1.84	129	trace →
15	Muse Spark 1.3 (xhigh)	90.8%	$1.51	217	trace →
16	GPT 5.5 (high)	90.0%	$3.54	36	trace →
17	GPT-5.6 Sol	88.8%	$1.14	17	trace →
18	Muse Spark 1.1 (xhigh)	87.8%	$0.24	108	trace →
19	Claude Sonnet 4.6	86.6%	$33.62	628	trace →
20	Gemini 3.6 Flash	82.9%	$3.15	98	trace →
21	Claude Opus 4.7 (xhigh)	82.9%	$2.60	89	trace →
22	GPT 5.5	80.8%	$1.09	19	trace →
23	Gemini 3 Flash	68.9%	$0.35	64	trace →
24	Gemini 3.1 Pro	66.6%	$0.84	45	trace →
25	Claude Opus 4.7	40.9%	$0.71	46	trace →
26	Claude Haiku 4.5	38.5%	$0.87	115	trace →
27	GPT 5.4	34.5%	$0.13	9	trace →
28	GPT 5.4 mini	26.1%	$0.04	8	trace →
29	Gemini 3.5 Flash	23.9%	$4.96	97	trace →
30	GPT 5 mini	13.8%	$0.01	15	trace →

Click a row to replay how that model rebuilt this program, or the model name to open its full run