SciCode Leaderboard

Science coding benchmark measuring AI ability to solve scientific computing tasks.

Data from Artificial Analysis

As of November 21, 2026, the top-scoring model on SciCode is Claude Fable 5.1 at 63.1%, followed by Gemini 4 Argon at 61.8% and Claude Fable 5 at 61.0%. 108 models have been evaluated on this benchmark.

Last updated: November 21, 2026

8 Ways to Use Fewer Tokens

Models

108

Best Score

63.1

Average

45.1

Std Dev

11.0

Categories
Computer Science and Programming
Sponsored The Grid | Spot Priced LLM API
The Grid | Spot Priced LLM API — Why pay list price when providers will bid for your API usage? Up to 80% off.
View spot prices
Provider
Model
Input $/M
Output $/M
SciCode
Actions
$10.000
$50.000
63.1
$2.000
$10.000
61.8
$10.000
$50.000
61.0
$2.000
$10.000
61.0
$0.326
$0.652
60.9
$0.240
$6.000
60.8
$0.180
$9.500
59.5
$0.070
$1.400
59.0
$1.250
$4.250
58.8
$1.000
$6.000
58.7
$1.250
$4.250
58.4
$0.100
$0.200
57.4
$2.000
$6.000
57.4
$0.750
$3.750
57.2
$2.000
$10.000
57.1
$0.750
$3.750
56.6
$2.000
$6.000
56.5
$5.000
$25.000
56.5
$5.000
$25.000
56.4
$5.000
$30.000
56.1
$1.000
$5.000
55.1
$2.000
$6.000
55.0
$2.000
$12.000
55.0
$0.100
$0.500
55.0
$5.000
$25.000
54.4
$2.000
$10.000
54.3
$0.700
$2.100
54.2
$0.900
$0.900
54.1
$1.500
$9.000
53.9
$0.200
$1.200
53.6
$0.750
$3.750
53.4
$0.135
$0.400
53.4
$2.000
$6.000
53.2
$0.375
$2.250
52.1
$0.040
$0.100
51.9
$0.040
$0.078
51.6
$0.000
$0.000
51.5
$0.100
$0.224
51.3
$0.007
$0.950
51.2
$0.111
$1.322
51.0
$0.480
$1.800
50.6
$0.014
$0.029
50.3
$3.000
$15.000
50.1
$0.000
$0.000
50.1
$0.450
$1.200
49.7
$0.216
$0.647
49.7
$1.475
$4.425
49.5
$0.132
$0.528
48.6
$0.150
$0.600
48.0
$0.600
$3.000
47.8
$0.050
$0.200
47.6
$0.100
$0.625
47.2
$0.230
$0.960
47.1
$1.000
$4.050
47.0
$0.035
$0.900
46.6
$1.000
$10.000
46.3
$0.320
$1.280
46.1
$3.000
$15.000
45.7
$0.100
$0.100
45.5
$0.390
$0.900
44.8
$0.899
$2.825
44.8
$0.060
$0.180
44.2
$0.200
$1.150
43.9
$0.250
$1.500
43.4
$0.250
$1.000
43.3
$1.100
$4.400
42.8
$0.250
$0.900
42.8
$0.042
$0.123
42.4
$1.000
$5.000
42.2
$0.021
$0.063
42.0
$0.450
$0.900
41.4
$0.300
$2.500
41.3
$0.400
$1.760
40.3
$1.500
$7.500
40.2
$0.050
$0.190
40.0
$0.260
$0.900
39.7
$0.250
$1.000
39.0
$0.125
$1.000
39.0
$0.017
$0.085
38.9
$0.040
$0.150
38.5
$0.300
$1.500
38.5
$0.550
$2.190
38.3
$0.270
$1.000
38.0
$0.250
$0.750
37.7
$0.050
$0.700
36.6
$0.300
$1.200
36.3
$0.120
$0.800
36.2
$0.064
$0.224
36.0
$0.030
$0.100
34.0
$0.200
$0.900
33.0
$0.400
$0.900
32.8
$0.440
$2.200
32.4
$0.020
$0.096
32.1
$0.200
$0.600
31.7
$0.060
$0.150
31.5
$0.080
$0.200
30.7
$0.060
$0.240
30.6
$0.080
$0.120
29.5
$0.075
$0.200
28.6
$0.200
$0.200
27.9
$0.100
$0.300
27.8
$0.150
$0.600
25.5
$0.200
$0.200
23.8
$0.080
$0.160
23.3
$0.100
$0.300
21.3
$0.150
$0.150
20.7
$0.050
$0.150
16.4
$0.350
$0.500
14.4

Pricing from OpenRouter. Benchmarks from Artificial Analysis.

About SciCode

Science coding benchmark measuring AI ability to solve scientific computing tasks.

This leaderboard shows all models with SciCode benchmark scores, ranked from highest to lowest. Pricing data is included to help you compare performance against cost.

Frequently Asked Questions

Science coding benchmark measuring AI ability to solve scientific computing tasks.
As of November 21, 2026, Claude Fable 5.1 leads the SciCode leaderboard with a score of 63.1. Rankings change as new models are released and evaluated.
Currently 108 models have been evaluated on SciCode, with an average score of 45.1 and standard deviation of 11.0.
Benchmark scores are updated when new evaluations are published by our data sources (Artificial Analysis and LayerLens). Pricing data is refreshed daily from OpenRouter.