Analog Design Bench

Benchmarking Agents on Hours-Long AMS Circuit Design TasksBenchmarking Agents on Hours-Long Analog and Mixed-Signal Circuit Design Tasks

Tasks: 50
Models: 24

Leaderboard

gpt-6-astra[max]

99%

Cost $3.60Out tok 29kSteps 30Agent time 23m

claude-opus-5-5[max]

97%

Cost $14.20Out tok 203kSteps 61Agent time 1h 03m

kimi-k3.1-pro[max]

88%

Cost $3.54Out tok 76kSteps 97Agent time 34m

claude-fable-5[max]

78%

Cost $31.56Out tok 108kSteps 37Agent time 1h 20m

deepseek-v4-1-flash-1017[max]

77%

Cost $0.29Out tok 371kSteps 124Agent time 56m

claude-opus-5[max]

69%

Cost $22.37Out tok 165kSteps 46Agent time 1h 23m

gpt-5-6-sol[max]

68%

Cost $14.63Out tok 74kSteps 148Agent time 1h 08m

gemini-4-argon[high]

65%

Cost $2.21Out tok 83kSteps 33Agent time 32m

gpt-5-5[xhigh]

61%

Cost $11.92Out tok 100kSteps 133Agent time 54m

gpt-5-6-terra[max]

50%

Cost $8.22Out tok 115kSteps 178Agent time 1h 26m

kimi-k3[max]

50%

Cost $5.52Out tok 133kSteps 75Agent time 1h 27m

deepseek-v4-1-flash-0910[max]

47%

Cost $0.30Out tok 347kSteps 91Agent time 1h 42m

deepseek-v4-pro-0813[max]

43%

Cost $0.37Out tok 293kSteps 108Agent time 1h 33m

mimo-v2-6-pro[thinking]

42%

Cost $0.38Out tok 217kSteps 86Agent time 1h 40m

glm-5-3[max]

33%

Cost $2.91Out tok 237kSteps 58Agent time 1h 43m

gpt-5-6-luna[max]

32%

Cost $1.39Out tok 133kSteps 216Agent time 1h 36m

qwen-3-8[xhigh]

30%

Cost $5.05Out tok 264kSteps 85Agent time 1h 44m

doubao-seed-evolving-0828[high]

26%

Cost $1.87Out tok 224kSteps 71Agent time 1h 50m

glm-5-3-flash[max]

19%

Cost $0.49Out tok 382kSteps 70Agent time 1h 56m

deepseek-v4-flash-0731[max]

16%

Cost $0.12Out tok 191kSteps 84Agent time 1h 55m

doubao-seed-2-1-pro[high]

14%

Cost $3.34Out tok 63kSteps 56Agent time 1h 57m

glm-5-2[max]

Cost $1.57Out tok 101kSteps 39Agent time 1h 52m

mimo-v2-5-pro[thinking]

Cost $2.66Out tok 332kSteps 78Agent time 1h 44m

nemotron-3-5-super-vl

Cost $0.52Out tok 186kSteps 78Agent time 1h 48m

Model
1gpt-6-astra[max]	99%	$3.60	29k	30	23m
2claude-opus-5-5[max]	97%	$14.20	203k	61	1h 03m
3kimi-k3.1-pro[max]	88%	$3.54	76k	97	34m
4claude-fable-5[max]	78%	$31.56	108k	37	1h 20m
5deepseek-v4-1-flash-1017[max]	77%	$0.29	371k	124	56m
6claude-opus-5[max]	69%	$22.37	165k	46	1h 23m
7gpt-5-6-sol[max]	68%	$14.63	74k	148	1h 08m
8gemini-4-argon[high]	65%	$2.21	83k	33	32m
9gpt-5-5[xhigh]	61%	$11.92	100k	133	54m
10gpt-5-6-terra[max]	50%	$8.22	115k	178	1h 26m
11kimi-k3[max]	50%	$5.52	133k	75	1h 27m
12deepseek-v4-1-flash-0910[max]	47%	$0.30	347k	91	1h 42m
13deepseek-v4-pro-0813[max]	43%	$0.37	293k	108	1h 33m
14mimo-v2-6-pro[thinking]	42%	$0.38	217k	86	1h 40m
15glm-5-3[max]	33%	$2.91	237k	58	1h 43m
16gpt-5-6-luna[max]	32%	$1.39	133k	216	1h 36m
17qwen-3-8[xhigh]	30%	$5.05	264k	85	1h 44m
18doubao-seed-evolving-0828[high]	26%	$1.87	224k	71	1h 50m
19glm-5-3-flash[max]	19%	$0.49	382k	70	1h 56m
20deepseek-v4-flash-0731[max]	16%	$0.12	191k	84	1h 55m
21doubao-seed-2-1-pro[high]	14%	$3.34	63k	56	1h 57m
22glm-5-2[max]	8%	$1.57	101k	39	1h 52m
23mimo-v2-5-pro[thinking]	8%	$2.66	332k	78	1h 44m
24nemotron-3-5-super-vl	5%	$0.52	186k	78	1h 48m
	0%25%50%75%100%