multiprocessio/dsq | ProgramBench

Commandline tool for running SQL queries against JSON, CSV, Excel, Parquet, and more.

542

Generated Behavioral Tests

97.4%

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)	97.4%	$31.41	305	trace →
2	Muse Wonder Full	96.7%	TBA	390	trace →
3	Claude Fable 5.1 (xhigh)	95.6%	$32.31	227	trace →
4	GPT 6 Astra (max)	94.8%	$13.19	66	trace →
5	Muse Wonder (xhigh)	93.5%	$42.30	382	trace →
6	Muse Spark 1.3 (max)	91.0%	$7.68	511	trace →
7	Claude Opus 5 (xhigh)	90.8%	$41.46	268	trace →
8	Muse Spark 1.3 (xhigh)	84.5%	$2.02	298	trace →
9	Claude Sonnet 4.6	80.3%	$32.92	610	trace →
10	GPT 5.5 (xhigh)	79.3%	$5.95	60	trace →
11	GPT-5.6 Sol (xhigh)	79.2%	$7.37	48	trace →
12	Gemini 3.5 Flash	78.0%	$8.37	231	trace →
13	Gemini 3.6 Flash	76.4%	$2.98	143	trace →
14	Muse Spark 1.2 (xhigh)	76.4%	$1.59	171	trace →
15	Claude Opus 4.6	76.4%	$14.97	351	trace →
16	Claude Opus 4.8 (xhigh)	72.5%	$11.09	104	trace →
17	GLM-5.2	71.0%	$22.32	267	trace →
18	Claude Opus 4.7 (xhigh)	68.5%	$6.13	105	trace →
19	Claude Opus 4.7	65.9%	$3.28	98	trace →
20	GPT-5.6 Sol	64.9%	$1.17	20	trace →
21	GPT 5.5 (high)	62.5%	$2.97	44	trace →
22	GPT 5.5	54.2%	$1.27	23	trace →
23	Gemini 3.1 Pro	50.6%	$1.64	100	trace →
24	GPT 5.4	49.3%	$0.54	24	trace →
25	Claude Haiku 4.5	30.3%	$0.77	134	trace →
26	GPT 5 mini	25.3%	$0.04	27	trace →
27	GPT 5.4 mini	10.1%	$0.03	8	trace →
28	Gemini 3.7 Flash	8.3%	$1.95	157	trace →
29	Muse Spark 1.1 (xhigh)	5.0%	$1.02	246	trace →
30	Gemini 3 Flash	0.0%	$0.38	118	trace →

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