Public Benchmark Registry

Comprehensive, error-free test results — download instead of running your own

Every entry here was produced with zero API errors and represents the most comprehensive test run (highest question count) for that model. Use these results to skip costly re-testing.

Models Registered

59

Avg Accuracy

70.0%

Total Questions Run

39,686

Last Updated

10/16/2026

RankModelProviderAccuracyQuestionsCorrect / TotalSpeedEnergyCostGDPRTested

neural-chat:latest

3919 MB

ollama
95.9%
706
677/706
0.11s/q
215.9
$3.0075
Yes
10/16/2026

zephyr:latest

3919 MB

ollama
89.4%
706
631/706
0.12s/q
195.2
$3.0287
Yes
10/16/2026

ibm/granite-3.2-8b

4714 MB

lmstudio
88.0%
50
44/50
0.85s/q
22.0
$0.2166
Yes
3/31/2026
4

llama3:latest

4445 MB

ollama
87.8%
706
620/706
1.20s/q
16.5
$2.5006
Yes
3/21/2026
5

mistral:7b

4170 MB

ollama
87.8%
706
620/706
1.23s/q
17.1
$2.9158
Yes
3/21/2026
6

llava-phi3:latest

2791 MB

ollama
87.5%
706
618/706
0.08s/q
378.6
$2.9911
Yes
10/16/2026
7

llava-phi3:3.8b

2791 MB

ollama
87.5%
706
618/706
0.07s/q
428.1
$2.9911
Yes
10/16/2026
8

mistral:latest

4170 MB

ollama
87.5%
706
618/706
1.42s/q
14.8
$2.9158
Yes
3/21/2026
9

llava:7b

4514 MB

ollama
87.0%
706
614/706
0.12s/q
162.1
$2.9511
Yes
10/17/2026
10

qwen2.5-coder:7b

4466 MB

ollama
86.3%
706
609/706
1.22s/q
15.8
$2.6350
Yes
3/21/2026
11

llava-llama-3-8b-v1_1

5289 MB

lmstudio
86.0%
50
43/50
0.85s/q
19.2
$0.2140
Yes
3/31/2026
12

bakllava:7b

4514 MB

ollama
85.8%
706
606/706
0.10s/q
199.1
$2.9724
Yes
10/17/2026
13

qwen2.5-coder:3b

1841 MB

ollama
85.0%
706
600/706
1.42s/q
32.5
$2.6350
Yes
3/21/2026
14

qwen2.5-coder:1.5b

940 MB

ollama
82.6%
706
583/706
1.11s/q
79.1
$2.6350
Yes
3/21/2026
15

llama3.1:8b

4693 MB

ollama
81.9%
706
578/706
1.05s/q
16.6
$2.5006
Yes
3/21/2026
16

llama3.1:latest

4693 MB

ollama
81.3%
706
574/706
0.22s/q
77.9
$2.5006
Yes
10/17/2026
17

minicpm-v:latest

5220 MB

ollama
81.0%
706
572/706
0.20s/q
77.6
$2.4867
Yes
10/17/2026
18

Gemma2:latest

5191 MB

ollama
80.0%
706
565/706
1.42s/q
10.8
$2.6341
Yes
3/21/2026
19

qwen2.5-coder:32b

18932 MB

ollama
79.5%
706
561/706
3.07s/q
1.4
$2.6350
Yes
3/21/2026
20

gemma3:1b

778 MB

ollama
79.3%
706
560/706
2.03s/q
50.2
$2.6281
Yes
3/22/2026
21

ibm/granite-4-micro

lmstudio
78.8%
706
556/706
0.11s/q
719.5
$2.6418
Yes
3/21/2026
22

ManishThota/llava_next_video:latest

4487 MB

ollama
78.6%
706
555/706
0.10s/q
183.8
$2.9982
Yes
10/17/2026
23

gemma3:270m

278 MB

ollama
78.0%
50
39/50
1.35s/q
207.1
$0.1924
Yes
3/31/2026
24

gemini-1.5-pro

base44
77.2%
706
545/706
1.17s/q
66.2—
No
3/21/2026
25

llava-llama3:latest

5289 MB

ollama
77.2%
706
545/706
0.25s/q
59.4
$2.5006
Yes
10/17/2026
26

llava-llama3:8b

5289 MB

ollama
77.2%
706
545/706
0.24s/q
61.9
$2.5006
Yes
10/17/2026
27

claude-3-5-sonnet-20241022

base44
77.1%
706
544/706
1.16s/q
66.3—
No
3/21/2026
28

gpt-4o-mini

base44
76.9%
706
543/706
1.15s/q
67.0—
No
3/21/2026
29

gpt-4o

base44
76.8%
706
542/706
1.13s/q
68.0—
No
3/21/2026
30

claude-3-5-haiku-20241022

base44
76.8%
706
542/706
1.14s/q
67.5—
No
3/21/2026
31

gemini-2.0-flash-exp

base44
76.5%
706
540/706
1.17s/q
65.5—
No
3/21/2026
32

google/gemma-3-4b

lmstudio
75.2%
706
531/706
0.11s/q
684.7
$2.6281
Yes
3/21/2026
33

qwen2.5:3b

1841 MB

ollama
74.4%
706
525/706
1.45s/q
27.9
$2.6350
Yes
3/21/2026
34

qwen2.5:14b

8572 MB

ollama
73.8%
706
521/706
2.13s/q
4.0
$2.6350
Yes
3/21/2026
35

llama3.2-vision:11b

7454 MB

ollama
71.2%
706
503/706
1.81s/q
5.3
$2.4935
Yes
3/21/2026
36

llava:13b

7640 MB

ollama
67.6%
706
477/706
0.48s/q
18.5
$2.9982
Yes
10/17/2026
37

llama3.2:1b

1260 MB

ollama
62.5%
706
441/706
0.19s/q
260.9
$2.6065
Yes
10/16/2026
38

qwen2.5:0.5b

379 MB

ollama
60.2%
706
425/706
0.17s/q
936.9
$2.6350
Yes
10/16/2026
39

qwen2.5-coder:14b

8572 MB

ollama
59.6%
706
421/706
2.16s/q
3.2
$2.6350
Yes
3/21/2026
40

qwen2.5-coder:0.5b

379 MB

ollama
58.6%
706
414/706
3.03s/q
51.1
$2.6350
Yes
3/22/2026
41

tinyllama:1.1b

608 MB

ollama
54.1%
706
382/706
0.08s/q
1073.6
$3.2147
Yes
10/16/2026
42

llama3.2:3b

1926 MB

ollama
51.4%
706
363/706
0.28s/q
95.1
$2.6065
Yes
10/16/2026
43

llama3.2:latest

1926 MB

ollama
51.4%
706
363/706
0.29s/q
92.5
$2.6065
Yes
10/16/2026
44

moondream:latest

1658 MB

ollama
51.3%
706
362/706
0.08s/q
391.3
$3.0682
Yes
10/16/2026
45

qwen3:latest

4983 MB

ollama
51.3%
706
362/706
0.19s/q
52.8
$2.5009
Yes
10/17/2026
46

qwen3:1.7b

1296 MB

ollama
51.3%
706
362/706
1.34s/q
29.5
$2.5573
Yes
3/21/2026
47

glm4:latest

5203 MB

ollama
51.3%
706
362/706
0.18s/q
54.2
$2.4582
Yes
10/17/2026
48

vicuna:latest

3649 MB

ollama
51.3%
706
362/706
0.08s/q
182.7
$3.2029
Yes
10/16/2026
49

qwen3-vl:latest

5856 MB

ollama
51.3%
706
362/706
0.30s/q
29.6
$2.5241
Yes
10/17/2026
50

qwen3:8b

4983 MB

ollama
51.3%
706
362/706
1.44s/q
7.1
$2.5573
Yes
3/21/2026
51

qwen3-vl:2b

1802 MB

ollama
51.3%
706
362/706
0.28s/q
101.5
$2.6059
Yes
10/16/2026
52

qwen3:0.6b

498 MB

ollama
51.3%
706
362/706
0.17s/q
620.6
$2.5009
Yes
10/16/2026
53

qwen3.8:27b

16920 MB

ollama
51.3%
706
362/706
5.15s/q
0.6
$2.7021
Yes
10/17/2026
54

openai/gpt-oss-20b

lmstudio
51.3%
706
362/706
1.05s/q
48.7
$2.9732
Yes
3/21/2026
55

lama3-8b-hikikomori-v0.3

lmstudio
51.0%
706
360/706
0.44s/q
117.2
$3.0750
Yes
3/22/2026
56

functiongemma:270m

287 MB

ollama
51.0%
706
360/706
2.19s/q
81.1
$2.6825
Yes
3/21/2026
57

llama2:latest

3650 MB

ollama
49.3%
706
348/706
0.11s/q
117.7
$3.0617
Yes
10/16/2026
58

mapler/llava-mistral:latest

4766 MB

ollama
48.7%
706
344/706
0.10s/q
100.4
$3.0787
Yes
10/17/2026
59

orca-mini:latest

1888 MB

ollama
48.7%
706
344/706
0.08s/q
319.9
$3.1826
Yes
10/16/2026