Perseus
Perseus

The Model Leaderboard

Compare reliability, extraction quality, hallucination rates, latency, and cost across all reviewed models.
Explore the full analytics table and identify the best model profile for your workload.

Best reliability

100.00%

Fastest average latency

1.36s

Lowest average cost

$0.01

RankModel
1Claude 4.5 Sonnet - high-thinking - 32768 tokens100.00.670.70.920.860.880.860.740.840.00.00.00.00.00.00.00.00.00.00.00.00.033.1520.0780402893018:32:48135.961619148758256160
2Claude 4.5 Sonnet - medium-thinking - 22188 tokens99.950.660.70.910.860.890.860.740.840.00.00.00.00.00.050.00.00.00.00.00.00.035.4220.0780402992019:48:59138.971619148760260910
3Mistral Small 3.2 - SFT99.950.870.90.920.950.970.950.820.926.013.131.290.150.00.00.00.00.050.00.050.00.07.9330.011606242004:26:1614.032337984866590
4GPT 5.2 2025-12-11 - xhigh-thinking99.90.720.770.90.840.860.840.720.820.00.00.00.00.00.10.00.00.00.00.00.00.062.060.066894311298434:43:0833.07138854066267026008979
5Qwen 3 32B - SFT99.90.860.90.920.940.960.940.810.926.063.031.590.50.050.00.050.050.00.00.00.00.013.4120.011584230007:30:1122.431907124626390
6Gemma 3 27B Inst - SFT99.80.860.890.920.950.960.950.820.934.872.781.290.250.00.00.050.150.00.00.00.00.014.5730.011552261008:09:0925.231264175247340
7Claude 4.1 Opus - high-thinking - 16384 tokens99.750.660.690.910.870.890.870.740.860.050.00.00.00.00.00.00.250.00.00.00.00.060.4850.2280401301033:50:17439.431619148726207160
8Claude 4.5 Sonnet - minimal-thinking - 11606 tokens99.750.660.70.910.860.890.860.740.840.00.00.00.00.00.00.00.050.00.20.00.00.027.2960.0680402393015:16:13120.861619148748191910
9GPT 5.2 2025-12-11 - high-thinking99.650.690.730.880.850.880.850.70.830.00.00.00.00.00.050.00.050.00.250.00.00.015.4740.03689430360108:39:2432.83138854066092321210664
10Gemma 3 12B Inst - SFT99.650.850.880.910.940.960.940.810.935.263.081.590.60.20.00.10.250.00.00.00.00.011.390.011552262006:22:1819.631264175280900
11Qwen 3 8B - SFT99.650.850.880.910.930.950.940.80.926.213.531.591.040.050.10.150.10.00.00.00.00.07.490.011584228004:11:2514.031907124599360
12Claude 4 Sonnet - medium-thinking - 22188 tokens99.550.650.690.910.840.870.840.730.820.00.00.00.00.00.10.00.20.00.150.00.00.029.1080.0680402609016:17:03127.391619148752544300
13Qwen 3 14B - SFT99.550.860.890.920.950.960.950.810.925.462.981.540.350.00.050.250.10.050.00.00.10.08.3280.011584230004:39:3314.031907124622330
14Claude 4.1 Opus - medium-thinking - 11264 tokens99.50.660.690.910.870.890.870.730.850.00.00.00.00.00.250.00.250.00.00.00.00.060.3710.2280401345033:46:28446.011619148727085580
15GPT 5 2025-08-07 - high-thinking99.50.720.790.880.850.880.850.70.840.00.00.00.00.00.00.00.00.00.50.00.00.076.4750.066894306446042:46:5923.51138854066153048982464
16GPT 5 Mini 2025-08-07 - high-thinking99.50.660.710.870.780.830.780.680.760.00.00.00.00.00.00.00.00.00.50.00.00.061.7040.016894306376934:31:114.71138854066171217591616
17Phi 4 - SFT99.40.840.890.90.940.960.940.770.96.94.221.741.340.20.00.30.250.00.00.050.00.07.0260.011574226003:55:4911.231706304542180
18Claude 4 Sonnet - high-thinking - 32768 tokens99.360.660.690.910.840.870.840.730.830.050.00.00.00.00.00.00.550.00.050.00.050.021.7580.0580401931012:10:21106.911619148738892880
19Claude 4 Sonnet - low-thinking - 11606 tokens99.260.650.690.90.840.870.840.730.820.10.00.00.00.00.00.00.650.00.10.00.00.021.5820.0580401906012:04:25106.151619148738381640
20Qwen 3 4B Inst - SFT99.110.840.880.910.940.960.940.790.915.863.181.590.650.00.050.150.70.00.00.00.00.07.2940.011584229004:04:4914.031907124604830
21Claude 4.1 Opus - minimal-thinking - 1024 tokens99.060.660.70.910.870.90.870.730.860.00.00.00.00.00.050.00.890.00.00.00.00.043.8590.198040953024:32:12386.851619148719196670
22Claude 4.5 Sonnet - minimal-thinking - 1024 tokens99.010.660.70.910.860.890.860.740.840.10.00.00.00.00.00.00.00.00.990.00.00.020.0840.0580401708011:14:08100.171619148734399690
23Gemma 3 4B Inst - SFT99.010.840.880.890.940.950.940.760.910.683.772.230.840.450.10.40.450.00.00.050.00.08.2130.011552263004:35:3914.031264175303100
24Claude 4.1 Opus - low-thinking - 6144 tokens98.960.660.690.910.870.90.870.730.860.050.00.00.00.00.450.00.60.00.00.00.00.048.6230.280401090027:12:06407.571619148721959220
25GPT 5 2025-08-07 - medium-thinking98.660.70.770.870.850.880.850.690.830.00.00.00.00.00.00.050.00.01.290.00.00.048.9210.046894302224327:22:0623.43138854066072284517888
26Qwen 3 0.6B - SFT98.360.830.880.90.930.950.930.740.8910.433.382.530.60.450.00.151.040.20.10.150.20.05.6380.011584230003:09:1511.231907124630070
27Qwen 3 1.7B - SFT98.210.840.880.90.930.950.940.760.919.883.432.530.750.10.650.550.550.00.00.050.00.05.6290.011584228003:08:5611.231907124595040
28GPT 5.2 2025-12-11 - medium-thinking97.960.670.720.870.850.870.850.690.840.00.00.00.00.00.350.250.00.01.440.00.00.010.9610.02689429937506:07:5532.7213885406601780754548
29Claude 4 Sonnet - minimal-thinking - 1024 tokens97.870.650.690.90.850.880.850.730.840.10.00.00.00.00.00.051.940.00.150.00.00.017.3640.0580401344009:42:5189.161619148727057980
30Gemini 2.5 Flash - dynamic-thinking97.220.650.70.90.850.90.850.710.840.050.10.00.10.00.00.00.60.02.190.00.00.029.1530.017317348143616:18:341.39147372517010552892074
31Gemini 2.5 Flash - low-thinking - 12288 tokens97.170.660.70.90.850.910.860.720.840.10.00.00.00.00.00.00.50.02.330.00.00.041.6460.017317349155923:17:551.39147372517025193140282
32GPT 5 Mini 2025-08-07 - medium-thinking97.070.650.70.860.80.850.80.670.780.00.00.00.00.00.00.00.40.02.530.00.00.029.0550.016894300148016:15:164.68138854066032002980416
33Gemini 2.5 Flash - medium-thinking - 18432 tokens96.770.650.70.890.850.910.850.710.840.050.10.050.10.00.00.00.840.052.330.00.00.030.1240.017317346153716:51:091.38147372516975813095506
34Gemini 2.5 Flash - minimal-thinking - 6144 tokens96.480.650.70.90.860.910.860.720.840.10.00.00.00.00.00.00.60.02.930.00.00.041.8130.017317348155123:23:301.39147372517006193122813
35Gemini 2.5 Flash - high-thinking - 24576 tokens96.430.660.70.90.840.90.840.720.820.350.20.20.20.150.00.00.890.02.680.00.00.030.6970.017317348159617:10:231.39147372517013353213751
36GPT 5 2025-08-07 - low-thinking96.380.670.740.850.860.880.860.670.840.00.00.00.00.00.10.050.30.03.180.00.00.016.2020.02689429676409:03:5123.32138854065960811538752
37GPT 4.1 2025-04-1495.830.680.720.890.90.930.90.680.884.720.550.00.150.050.01.042.830.250.050.00.050.02.190.026895286001:13:3132.38138874205759850
38Claude 3.5 Sonnet V294.790.630.690.870.880.920.890.650.871.240.050.00.00.00.01.643.380.00.00.00.00.27.410.038010352004:08:4259.03161310677091890
39Gemini 2.5 Pro - medium-thinking - 21888 tokens94.690.670.70.910.850.880.850.740.830.150.00.00.00.00.00.050.00.05.260.00.00.018.4170.037317357194210:18:1125.61147372517184613911413
40Gemini 2.5 Pro - low-thinking - 11008 tokens93.990.670.710.910.850.880.850.730.830.00.00.00.00.00.00.00.00.06.010.00.00.016.8770.037317357187609:26:3025.6147372517182463777495
41Gemini 2.5 Pro - dynamic-thinking93.60.660.70.910.860.880.860.740.830.150.00.00.00.00.00.00.00.06.410.00.00.017.6830.037317358186109:53:3325.62147372517200823746963
42Claude 3.7 Sonnet - no-thinking93.350.640.710.870.870.910.870.670.852.530.00.00.00.00.00.653.770.152.090.00.00.06.6680.038010364003:43:4859.4161310677337260
43GPT 5.2 2025-12-11 - no-thinking93.350.570.60.690.860.880.860.540.852.090.00.00.00.00.150.352.230.03.920.00.00.02.9580.026894242001:39:1631.11138854064863460
44GPT 5.2 2025-12-11 - low-thinking92.30.650.70.860.850.870.850.680.830.10.00.00.00.00.50.050.350.06.80.00.00.07.4050.02689429517104:08:3232.6213885406594275343633
45Claude 4.5 Sonnet - no-thinking91.960.680.720.90.870.920.870.720.851.490.250.00.050.00.20.945.710.20.990.00.00.05.3250.038009366002:58:4559.45161290537372740
46Gemini 2.5 Pro - high-thinking - 32768 tokens91.560.670.70.910.850.880.850.740.830.150.00.00.00.00.00.00.00.08.440.00.00.018.2520.037317357192510:12:3925.61147372517189973876609
47Claude 3.5 Sonnet V190.670.620.70.860.860.90.860.650.844.221.290.00.350.00.23.435.310.20.150.050.00.07.3340.038010369004:06:1059.55161310677439550
48Claude 4 Sonnet - no-thinking89.820.690.720.910.880.910.880.740.861.690.50.00.050.00.01.099.090.00.00.00.00.04.9430.038010362002:45:5459.32161310677287000
49Claude 4.1 Opus - no-thinking89.030.670.710.90.90.940.90.740.891.990.10.00.00.00.01.549.430.00.00.00.00.018.7240.158010356010:28:30295.79161310747176060
50Claude 3 Opus87.790.660.70.890.820.860.820.70.818.244.820.32.730.050.02.387.551.540.890.10.20.015.4360.158010417008:38:07304.95161310728397830
51Claude 4.5 Haiku - high-thinking - 32768 tokens87.490.660.70.90.850.890.850.720.830.050.00.00.00.00.00.00.00.012.510.00.00.027.4010.0380404548015:19:4561.991619148791586740
52Claude 4 Opus - no-thinking87.140.670.710.910.90.940.90.740.892.330.050.00.00.00.152.3810.330.00.00.00.00.017.1080.158010359009:34:15296.13161310677221290
53Claude 4.5 Haiku - medium-thinking - 22188 tokens86.20.660.70.90.850.890.850.720.830.00.00.00.00.00.00.00.050.013.750.00.00.023.7040.0380403748013:15:3953.941619148775492490
54Gemini 2.5 Pro - minimal-thinking - 128 tokens85.90.680.730.880.880.930.880.720.871.440.00.00.00.00.00.51.840.011.770.00.00.03.5690.0173173428401:59:4625.314737251687692168106
55Claude 4.6 Sonnet - dynamic-thinking85.40.650.70.90.830.860.830.730.810.00.00.00.00.00.00.00.00.014.60.00.00.023.7470.0580281854013:17:05104.51616731937331140
56GPT 5.1 2025-11-13 - high-thinking84.260.760.820.890.750.850.750.720.730.00.00.00.00.00.10.00.00.015.640.00.00.053.7120.066894313470930:02:5423.66138854066297909484749
57GPT 5.1 2025-11-13 - medium-thinking82.420.730.790.890.780.850.780.720.760.00.00.00.00.00.050.00.250.017.280.00.00.028.7720.036894311159716:05:4723.62138854066262243216618
58Claude 4.5 Haiku - low-thinking - 11606 tokens80.440.650.690.890.850.890.850.710.820.050.00.00.00.00.00.00.10.019.460.00.00.017.0810.0280402747009:33:2143.861619148755329800
59GPT 5 Mini 2025-08-07 - low-thinking80.440.660.70.860.80.840.80.680.780.00.00.00.00.00.00.01.390.018.120.00.050.015.5790.01689429853308:42:554.67138854065996311073344
60GPT 5.1 2025-11-13 - no-thinking75.970.620.660.760.850.880.850.590.831.940.60.050.20.00.051.242.780.6519.320.00.00.03.8180.016894280002:08:0922.99138854065630450
61Claude 4.5 Haiku - minimal-thinking - 1024 tokens75.320.650.690.890.850.890.850.710.830.150.00.00.00.00.00.00.10.024.580.00.00.017.9860.0280402795010:03:4344.341619148756286770
62Claude 4.6 Sonnet - no-thinking68.370.660.710.870.840.90.840.690.820.60.450.00.050.00.00.150.350.0531.080.00.00.04.4620.038011372002:29:4659.64161330817494960
63GPT 5 2025-08-07 - minimal-thinking66.630.630.680.790.870.90.870.610.851.640.050.050.00.00.01.141.040.430.790.00.00.04.950.016894289002:46:0823.18138854065821860
64Magistral Medium 1.264.750.640.690.870.740.840.740.690.720.20.150.00.10.051.490.050.30.033.320.00.40.0104.7710.0370483602058:36:4864.661419398672533710
65GPT 5.1 2025-11-13 - low-thinking60.820.690.740.860.760.840.760.70.730.10.050.00.00.050.050.01.840.1537.140.00.050.012.1560.02689430948406:48:0223.5813885406622295974434
66Claude 3.5 Haiku60.530.650.720.860.720.820.720.640.74.522.040.00.60.00.256.88.592.7821.650.00.550.07.210.018010439004:02:0116.44161310678848120
67Mistral Medium 357.70.660.730.860.620.70.620.640.612.5312.260.051.540.00.00.794.9713.2126.220.050.20.06.9710.017048351003:53:597.09141939867072980
68GPT 4.1 Mini 2025-04-1457.00.750.790.890.850.890.850.690.834.671.890.20.70.00.05.115.811.6430.540.00.890.03.6850.016895300002:03:416.52138874206036100
69Gemini 2.5 Flash - no-thinking56.850.670.710.90.810.880.820.710.793.630.30.050.150.00.034.610.70.457.450.00.00.01.3640.017317360000:45:476.24147372517258110
70Mistral Large 355.560.620.70.850.590.690.590.650.555.965.360.22.230.00.554.724.423.6730.980.351.090.05.1910.017048372002:54:148.22141939867497680
71Gemini 2.5 Flash-Lite - no-thinking55.360.620.660.880.330.40.330.60.311.8224.431.4414.80.150.53.0822.4914.49.630.114.40.01.5070.017317663000:50:341.641473725113350220
72Claude 4.5 Opus - no-thinking53.970.70.730.920.830.890.830.760.810.30.050.00.00.00.00.250.890.044.890.00.00.05.2430.058010370002:55:5899.31161310677460500
73Gemini 2.5 Flash-Lite - minimal-thinking - 512 tokens52.190.640.690.880.390.440.390.610.3610.8727.951.0413.80.10.254.0213.6113.1120.660.27.940.02.5640.01731763843501:26:031.62147372511285400876635
74Gemini 2.0 Flash51.290.650.690.890.780.880.780.690.764.120.30.00.050.00.4541.412.530.254.070.00.050.01.80.017317347001:00:252.63147372516986930
75Gemini 2.5 Flash-Lite - dynamic-thinking50.350.690.730.90.610.690.610.70.570.40.650.00.350.00.00.31.640.1547.270.01.040.012.8750.017317377376707:12:101.41147372517598077587222
76Ministral 3 8B50.250.610.670.860.620.770.620.590.598.495.070.052.680.03.286.8511.324.8224.980.252.230.12.0180.017048328001:07:442.23141939866607200
77Gemini 2.5 Flash-Lite - high-thinking - 24576 tokens49.950.70.750.890.620.690.620.690.580.70.40.00.150.050.00.51.640.147.720.00.940.010.3430.017317379411405:47:101.41147372517634958285748
78Gemini 2.5 Flash-Lite - low-thinking - 8534 tokens49.650.680.730.90.610.690.610.710.580.650.350.00.10.00.00.31.640.448.010.00.940.010.3880.017317379398205:48:411.41147372517636308020527
79Gemini 2.5 Flash-Lite - medium-thinking - 16556 tokens47.920.70.740.90.620.70.620.710.590.650.30.00.20.00.00.451.740.1549.650.01.090.010.4640.017317377407405:51:141.41147372517596648205794
80Claude 4.6 Opus - no-thinking45.780.680.730.910.730.830.730.760.70.150.150.00.00.00.250.10.10.053.770.00.00.06.5340.058011507003:39:20106.211613308110216580
81Mistral Small 4 - high-thinking45.280.660.710.880.670.750.670.70.641.092.190.00.20.00.254.972.981.9444.740.11.340.011.9860.0170602175006:42:194.761421815443806870
82Claude 4.5 Haiku - no-thinking41.660.640.670.880.80.870.80.70.771.940.40.00.150.10.01.192.190.754.320.00.00.03.5880.018010374002:00:2519.89161310677523600
83GPT 5 Mini 2025-08-07 - minimal-thinking41.510.70.740.870.670.750.680.670.643.233.870.051.740.050.01.743.771.8451.490.00.990.05.0530.016894322002:49:374.77138854066487130
84Ministral 3 14B39.370.630.670.870.540.710.550.630.494.425.160.01.940.00.2514.47.12.7336.990.11.040.03.0890.017048317001:43:402.97141939866387500
85GPT 4o 2024-11-2039.030.670.710.870.760.870.760.670.732.481.790.10.60.050.791.691.091.0956.410.050.150.02.0550.026895303001:08:5840.82138874206104880
86Claude 4.6 Opus - dynamic-thinking38.330.680.720.910.690.770.690.720.670.00.00.00.00.015.290.00.00.046.380.00.00.019.0270.0780281231010:38:40142.821616731924794990
87GPT 5 Nano 2025-08-07 - high-thinking37.340.670.720.870.70.770.70.680.660.00.60.00.30.02.535.367.851.9445.280.06.80.090.4460.0168944041006950:35:581.021388540681351320279232
88Mistral Medium 3.135.250.650.710.870.670.740.670.660.641.395.560.00.150.00.08.792.3818.3736.840.150.250.07.110.017048364003:58:397.15141939867338230
89Mistral Small 4 - no-thinking34.810.510.540.70.680.790.680.540.651.944.320.00.750.00.05.464.075.6151.890.10.60.01.8830.017060309001:03:122.51142181546231010
90Gemini 2.0 Flash-Lite32.920.660.690.90.70.860.70.70.672.090.70.050.150.00.046.482.481.2916.980.00.150.12.6760.017317377001:29:481.33147372517597630
91Claude 3 Haiku32.470.60.650.830.370.570.370.590.357.39.631.193.080.890.0510.5317.6830.8322.050.553.180.05.2120.018010502002:54:555.31613106710113840
92Claude 3 Sonnet30.040.650.690.890.660.770.660.690.632.985.960.152.880.00.04.7710.924.3750.890.357.850.06.9850.038010398003:54:2760.43161310678023930
93GPT 5 Nano 2025-08-07 - medium-thinking28.90.660.710.860.680.750.680.670.630.10.790.00.450.011.5216.787.052.0934.760.04.920.0552.4350.016894668499529:20:031.2313885406134445610060480
94GPT 4o Mini 2024-07-1826.220.650.670.890.590.740.60.630.573.084.270.151.640.355.4119.5126.2216.4418.720.017.630.06.5060.016895322003:38:232.47138874206489700
95GPT 4.1 Nano 2025-04-1410.920.650.70.840.650.850.670.590.591.843.480.351.640.31.6957.7515.2910.539.930.997.990.153.2210.016895311001:48:051.64138874206267090
96GPT 5 Nano 2025-08-07 - low-thinking7.750.640.670.790.840.880.840.610.790.050.20.00.050.029.7947.623.872.1910.580.01.590.0523.2230.0168941142117412:59:301.611388540622994452365248
97GPT 5 Nano 2025-08-07 - minimal-thinking3.030.770.80.870.470.720.490.550.440.40.790.050.350.02.8847.0718.2716.8825.520.456.460.12.9750.016894386001:39:511.01138854067770700
Scroll horizontally to explore all analytics. Click any column header to sort.

Start free, scale with confidence

Launch your first graph-powered agent flow quickly, then scale to enterprise throughput with reliability, support, and infrastructure options designed for production teams.
DevelopREST & Python SDK
ExecuteServerless compute
ScaleEnterprise ready
MonitorConsole & alerts