Browse canonical models across providers with performance and coverage highlights.
Showing 1-24 of 36 models
Qwen3
Providers
34
Variants
85
Avg speed
27.18 t/s
First token
24.52 s
Tests
7,255
QwQ
13
19
28.86 t/s
26.73 s
6,588
glm-4
21
50
36.81 t/s
1.25 s
5,185
DeepSeek-V3
65
96
721.22 t/s
3.36 s
950
qwen2
11
56
59.67 t/s
1.23 s
490
Qwen2.5
45
50.52 t/s
1.34 s
405
test
26
138.96 t/s
2.40 s
194
deepseek-v3.1
16
3269.13 t/s
2.03 s
175
gpt-oss
15
296.95 t/s
4.08 s
165
glm-4.5
14
20
49.80 t/s
9.95 s
145
llama-4
8
303.97 t/s
0.82 s
105
qwen3-coder
12
97.16 t/s
2.41 s
70
qwen3-next
7
145.19 t/s
1.14 s
qwen2.5-coder
3
51.20 t/s
0.68 s
60
qwen2.5-vl
40.03 t/s
1.56 s
55
Qwen3-VL
4
5
64.27 t/s
12.99 s
40
DeepSeek-V3.2
26.73 t/s
2.06 s
glm-4.1v-thinking
2
108.16 t/s
7.48 s
internlm2
68.06 t/s
0.61 s
30
llama-3.3
770.59 t/s
0.46 s
qwen2-vl
1
69.34 t/s
0.85 s
phi-4
53.44 t/s
1.03 s
GLM-4.6
84.59 t/s
2.28 s
10
deepseek-v2
15.51 t/s
0.96 s