LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Best DeepSeek model overall

DeepSeek V4.1 Flash 1017 leads
39.5
models ranked
13
lead over #2
+3.5

Overall index · read October 21

As of October 21, 2026, DeepSeek V4.1 Flash 1017 leads 13 models that qualify, with 39.5 on the overall index, ahead of DeepSeek V4 Pro 0813 (36.0) and DeepSeek V4 Flash Vision Exp (34.8).

13 models, higher is better

Embed this data
Best DeepSeek model overall
Rank Model Overall index Evidence Coding index Output $/Mtok
1
39.5
83.3 $0.60
2
36.0
54.6 $1.98
3
34.8
37.6 $0.6468
4
34.3
Not benchmarked $1.32
5
30.4
35.3 $1.9105
6
24.7
77.2 $0.525
7
24.4
54.6 $0.182
8
DeepSeek V3.2 DeepSeek
21.5
54.4 $0.42
9
16.6
55.0 $0.41
10
14.8
Not benchmarked $1.00
11
DeepSeek V3.1 DeepSeek
13.7
Not benchmarked $0.95
12
R1 0528 DeepSeek
13.1
71.4 $2.15
13
DeepSeek V3 0324 DeepSeek
9.7
42.0 $1.14

How this list is built

Generated from the dataset on every refresh. Nobody picks the order by hand.

Every DeepSeek model on the overall index, ordered by the Artificial Analysis Intelligence Index, highest first. The index is a composite of evaluations Artificial Analysis runs itself, published on a 0 to 100 scale.

Other lists

Rankings beyond coding · Methodology