facebookresearch/fastText

Library for fast text representation and classification.

312

Generated Behavioral Tests

98.1%

Best Score

Claude Opus 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)

25 runs

#	Model	Score	Cost	Calls
1	Claude Opus 5 (xhigh)	98.1%	$23.54	149	trace →
2	Claude Opus 4.8 (xhigh)	96.5%	$10.96	79	trace →
3	Muse Spark 1.3 (max)	92.0%	$5.21	299	trace →
4	Muse Spark 1.3 (xhigh)	87.5%	$0.98	202	trace →
5	Muse Spark 1.1 (xhigh)	84.3%	$0.27	190	trace →
6	GPT 5.5 (xhigh)	80.4%	$8.25	80	trace →
7	GLM-5.2	78.2%	$25.55	178	trace →
8	GPT-5.6 Sol (xhigh)	77.6%	$5.34	48	trace →
9	Gemini 3.7 Flash	77.2%	$2.51	135	trace →
10	Claude Opus 4.6	75.6%	$15.58	242	trace →
11	Muse Spark 1.2 (xhigh)	71.5%	$0.08	43	trace →
12	Claude Sonnet 4.6	67.9%	$13.78	294	trace →
13	GPT-5.6 Sol	63.1%	$0.82	16	trace →
14	GPT 5.5 (high)	62.8%	$2.49	27	trace →
15	GPT 5.5	61.5%	$1.02	14	trace →
16	Gemini 3.6 Flash	52.2%	$4.51	116	trace →
17	Claude Opus 4.7	38.8%	$1.59	57	trace →
18	GPT 5.4	37.5%	$0.24	9	trace →
19	Claude Opus 4.7 (xhigh)	26.9%	$2.35	60	trace →
20	Gemini 3 Flash	15.4%	$0.22	47	trace →
21	Claude Haiku 4.5	15.3%	$1.02	106	trace →
22	GPT 5 mini	8.3%	$0.01	8	trace →
23	Gemini 3.5 Flash	5.4%	$3.34	107	trace →
24	Gemini 3.1 Pro	5.4%	$1.54	63	trace →
25	GPT 5.4 mini	0.3%	$0.02	11	trace →

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