Models tracked
—
Across the selected period
Community telemetry
See how models perform in the real world.
Usage, speed, value, and intelligence scores—grounded in how CodingFleet users actually work, not marketing claims.
Tokens served
—
Prompt and completion tokens
Fastest response
—
Output tokens per second
Community favorite
—
Models with 10+ votes
Token usage over time
Model usage across CodingFleet.
Loading usage data…
Model leaderboard
Sort any metric, then narrow the list without reloading.
| Model | Speed | Usage | Credits | User score | Intelligence |
|---|---|---|---|---|---|
1
DeepSeek V4 Pro
|
82.6 tok/s | 13.8B | 2–4 | 0.7 | 36.0 |
|
2
DeepSeek V4 Flash
|
155.8 tok/s | 13.3B | 1 | 0.8 | 40.8 |
3
GPT-5.6 Luna
|
82.3 tok/s | 8.1B | 0.5–0.9 Unlimited | 0.4 | 37.3 |
4
Claude Haiku 4.5
|
115.5 tok/s | 5.2B | 2–3 | 0.1 | — |
|
5
Claude Sonnet 4.5
|
56.0 tok/s | 4.6B | 7–8 | 0.5 | — |
|
6
DeepSeek V4.1 Flash
|
181.3 tok/s | 3.4B | 1 Unlimited | 0.8 | 39.5 |
|
7
Claude Sonnet 4.6
|
57.1 tok/s | 3.2B | 9–10 | 1.0 | 36.8 |
|
8
GPT-5.2
|
47.8 tok/s | 2.9B | 4–8 | 0.8 | 43.3 |
|
9
Claude Opus 4.6
|
48.4 tok/s | 2.6B | 15–17 | 1.0 | 38.8 |
|
10
GPT-5.6 Terra
|
70.3 tok/s | 2.3B 1 | 5–7.5 | 0.8 | 42.1 |
|
11
GPT-5.4
|
65.1 tok/s | 2.3B 1 | 8–12 | 0.9 | 53.1 |
|
12
GPT-5 Mini
|
63.6 tok/s | 2.2B 2 | 1–2 | 0.7 | 31.6 |
13
Gemini 3 Flash
|
141.1 tok/s | 2.1B | 1 | 0.2 | 17.9 |
|
14
Claude Opus 4.5
|
57.0 tok/s | 2.0B | 15–17 | — | 41.9 |
|
15
GPT-5.4 Mini
|
119.1 tok/s | 1.8B | 1–2 | 0.3 | 40.9 |
|
16
Claude Sonnet 5
|
89.7 tok/s | 1.6B 1 | 6.8–10 | — | 38.2 |
|
17
Claude Opus 4.8
|
66.0 tok/s | 1.5B 2 | 15–17 | — | 41.8 |
|
18
GPT-5
|
49.9 tok/s | 1.0B | 3–5 | 1.0 | 35.3 |
|
19
Claude Opus 4.7
|
67.4 tok/s | 1.0B 1 | 15–17 | — | 55.0 |
|
20
GPT-5.1
|
69.9 tok/s | 760.8M 2 | 3–5 | 0.9 | 37.5 |
|
21
GPT-5.5
|
58.8 tok/s | 667.2M 2 | 14–24 | — | 56.3 |
|
22
Gemini 3.1 Pro
|
89.5 tok/s | 657.5M 3 | 7 | 1.0 | 29.7 |
|
23
GPT-5.2 Codex
|
50.5 tok/s | 654.6M 3 | 5–9 | 0.8 | 41.2 |
24
Qwen 3.8 Flash
|
57.1 tok/s | 558.2M 6 | 1 Unlimited | — | — |
|
25
GPT-5.1 Codex Mini
|
101.4 tok/s | 555.0M 6 | 1 | 0.7 | 31.3 |
|
26
GPT-5.3 Codex
|
59.0 tok/s | 497.3M 8 | 5–9 | — | 45.5 |
|
27
Gemini 3 Pro
|
77.3 tok/s | 410.9M 9 | 4 | 1.0 | 40.6 |
|
28
GPT-5.6 Sol
|
45.5 tok/s | 391.9M 9 | 11–18 | 1.0 | 47.0 |
|
29
Qwen 3.7 Flash
|
111.8 tok/s | 371.9M 9 | 1 Unlimited | — | — |
|
30
DeepSeek V3.2
|
26.8 tok/s | 316.7M 6 | 1 | — | 25.1 |
31
MiniMax M2.7
|
62.9 tok/s | 287.0M 11 | 1 | — | 22.8 |
|
32
Claude Opus 5
|
69.3 tok/s | 266.7M 11 | 15–22 | — | 50.8 |
33
GLM 5.1
|
42.5 tok/s | 249.9M 11 | 2 | — | 41.0 |
|
34
Claude 4 Sonnet
|
55.7 tok/s | 236.2M 18 | 6–7 | 0.9 | 29.8 |
|
35
Claude Fable 5 Thinking
|
68.8 tok/s | 230.5M 12 | 36 | — | 62.1 |
|
36
Gemini 2.5 Flash
|
188.8 tok/s | 205.1M 4 | 1 | 0.3 | 14.2 |
|
37
GPT-4.1
|
45.6 tok/s | 204.2M 8 | 6 | — | 19.6 |
|
38
MiniMax M3
|
68.9 tok/s | 198.2M 12 | 1 | — | 29.2 |
|
39
Qwen 3.6 Plus
|
50.2 tok/s | 179.1M 13 | 1 Unlimited | — | 27.0 |
|
40
Qwen 3.7 Max
|
45.9 tok/s | 169.5M 14 | 3 | — | 46.7 |
|
41
GLM 5.2
|
60.1 tok/s | 169.3M 14 | 2 | — | 33.7 |
|
42
GLM 5.3 Flash
|
56.2 tok/s | 166.8M 14 | 1 Unlimited | — | 41.8 |
43
Kimi K2.5
|
42.9 tok/s | 145.2M 14 | 1 | — | 36.0 |
44
Hy3
|
72.9 tok/s | 141.7M 14 | 1 Unlimited | — | 25.3 |
|
45
Gemini 3.8 Flash
|
114.0 tok/s | 141.2M 14 | 1.5–2.5 | — | 40.9 |
|
46
Gemini 3.5 Flash
|
143.2 tok/s | 137.1M 15 | 4 | 1.0 | 52.0 |
47
Nemotron 3.5 Lightning
|
82.5 tok/s | 128.2M 15 | 0 | — | 12.9 |
|
48
Qwen 3.6 Flash
|
165.0 tok/s | 121.2M 15 | 1 | — | — |
|
49
GLM 5
|
57.6 tok/s | 120.2M 15 | 1 | — | 40.6 |
|
50
Kimi K2.7 Code
|
67.8 tok/s | 118.6M 16 | 1 | — | 25.8 |
51
Muse Spark 1.3 Contributor
|
111.2 tok/s | 116.5M 16 | 1 Unlimited | — | — |
|
52
Gemini 2.5 Pro
|
93.4 tok/s | 102.5M 13 | 4 | — | 25.9 |
|
53
GPT-5.1 Codex Max
|
72.2 tok/s | 98.6M 16 | 4–5 | — | — |
|
54
Claude 4.1 Opus
|
38.9 tok/s | 98.4M 6 | 50–60 | — | 34.5 |
|
55
GLM 5 Turbo
|
54.7 tok/s | 86.2M 15 | 1 | — | 39.1 |
|
56
GPT-4.1 mini
|
47.8 tok/s | 84.2M 23 | 1 | — | 14.8 |
57
Grok-4.1 Fast
|
112.7 tok/s | 80.1M 14 | 1 | — | 17.0 |
58
MiMo V2.5 Pro
|
64.2 tok/s | 78.9M 14 | 1 | — | 26.0 |
|
59
Grok-4 Fast
|
156.2 tok/s | 76.7M 14 | 1 | 0.4 | 16.6 |
|
60
Gemini 3.6 Flash
|
160.6 tok/s | 73.7M 14 | 3 | — | 51.6 |
|
61
Kimi K2.6
|
51.8 tok/s | 72.7M 15 | 1 | — | 45.1 |
|
62
Muse Spark 1.2 Contributor
|
131.9 tok/s | 71.9M 15 | 1 Unlimited | — | — |
|
63
Qwen 3.5 Flash
|
95.1 tok/s | 68.0M 15 | 1 | — | — |
|
64
GLM 5V Turbo
|
68.0 tok/s | 66.5M 16 | 1 | — | 35.3 |
|
65
Qwen 3.5 Plus
|
57.6 tok/s | 63.5M 16 | 1 | — | — |
|
66
GLM 5.3
|
66.2 tok/s | 55.4M 17 | 2–3 | — | 44.8 |
|
67
Gemini 3.7 Flash
|
134.1 tok/s | 55.0M 17 | 2 | — | 39.1 |
|
68
Claude 3.5 Haiku
|
57.7 tok/s | 47.2M 47 | 1 | — | 12.2 |
|
69
Grok Build 0.1
|
125.7 tok/s | 41.5M 20 | 1 | — | — |
|
70
MiniMax M2.5
|
50.9 tok/s | 39.5M 21 | 1 | — | 34.5 |
|
71
Nemotron 3 Ultra
|
28.6 tok/s | 39.4M 21 | 0 | — | — |
72
Ling 3.0 Flash
|
239.5 tok/s | 35.2M 21 | 1 Unlimited | — | 24.9 |
|
73
Qwen 3.8 Max
|
35.6 tok/s | 33.2M 21 | 3 | — | 45.4 |
|
74
Kimi K3
|
32.6 tok/s | 30.3M 22 | 7 | — | 43.6 |
|
75
Claude Fable 5.1
|
71.7 tok/s | 26.3M 22 | 30–44 | — | 53.4 |
|
76
Laguna S 2.1
|
65.5 tok/s | 23.6M 23 | 0 | — | — |
|
77
Grok 4.6
|
54.7 tok/s | 21.5M 23 | 2–4 | — | 44.3 |
|
78
Qwen3 235B A22B
|
43.3 tok/s | 21.1M 20 | 1 | — | — |
|
79
MiMo V2 Pro
|
82.2 tok/s | 18.6M 24 | 1 | — | 41.4 |
|
80
Qwen 3.7 Plus
|
44.6 tok/s | 17.6M 24 | 1 | — | 25.2 |
|
81
Grok 4.5
|
54.3 tok/s | 17.3M 24 | 4 | — | 55.8 |
|
82
Qwen 3.6 Max Preview
|
38.5 tok/s | 16.8M 24 | 4 | — | 41.1 |
|
83
DeepSeek V3.2 Thinking
|
32.6 tok/s | 16.2M 4 | 1 | — | — |
|
84
Ox Alpha
|
29.7 tok/s | 15.1M 25 | 0 | — | — |
|
85
Grok-4.20 Beta
|
163.1 tok/s | 13.6M 26 | 4 | — | — |
|
86
GLM 4.6
|
45.2 tok/s | 13.0M 26 | 1 | — | 23.4 |
|
87
GPT-5 Pro
|
15.1 tok/s | 12.9M 26 | 70 | — | — |
|
88
Hy4 Preview
|
58.9 tok/s | 12.5M 26 | 1–2 | — | — |
|
89
Grok-4
|
40.9 tok/s | 10.1M 18 | 5 | — | 34.1 |
|
90
OpenAI o3
|
61.0 tok/s | 8.0M 42 | 4–5 | — | 31.1 |
91
Codestral (2508)
|
109.3 tok/s | 7.4M 27 | 1 | — | — |
|
92
Mistral Medium 3
|
69.1 tok/s | 7.3M 23 | 1 | — | 12.5 |
|
93
GLM 4.7
|
46.3 tok/s | 7.2M 26 | 1 | — | 34.5 |
|
94
GPT-6 Astra
|
30.4 tok/s | 6.4M 27 | 30–42 | — | 52.7 |
|
95
Mistral Medium 3.5
|
99.6 tok/s | 6.0M 27 | 2 | — | 14.2 |
|
96
GPT-5.2 Codex
|
66.3 tok/s | 5.2M 27 | 8 | — | 41.2 |
|
97
Muse Spark 1.3
|
143.7 tok/s | 3.5M 29 | 1–3 | — | 45.1 |
|
98
Llama 4 Maverick
|
55.8 tok/s | 3.5M 3 | 1 | — | 14.5 |
|
99
Kimi K2 Thinking
|
73.4 tok/s | 3.0M 31 | 1 | — | 33.5 |
|
100
Grok-4.20
|
175.7 tok/s | 2.8M 31 | 4 | — | 38.0 |
|
101
Qwen 3.8 27B
|
16.8 tok/s | 2.5M 32 | 0 | — | 33.7 |
|
102
Claude 3.7 Sonnet
|
54.4 tok/s | 1.8M 74 | 6–7 | — | 27.6 |
|
103
Gemini 2.0 Flash
|
138.3 tok/s | 1.8M 62 | 1 | — | 12.2 |
|
104
Qwen3 Coder
|
82.8 tok/s | 1.6M 23 | 2 | — | — |
|
105
Grok-4.3
|
110.3 tok/s | 1.4M 30 | 3 | — | 37.9 |
|
106
Muse Spark 1.2
|
145.9 tok/s | 1.1M 30 | 1 | — | 46.8 |
|
107
Kimi K2 (0905)
|
49.6 tok/s | 1.0M 27 | 1 | — | 19.7 |
|
108
Mistral Large 3
|
60.4 tok/s | 911.2K 29 | 1 | — | 4.1 |
|
109
Claude 4 Opus
|
25.3 tok/s | 389.5K 24 | 50–60 | — | 31.7 |
|
110
GPT-4o Mini
|
64.8 tok/s | 187.1K 78 | 1 | — | 6.7 |
|
111
GPT-4o
|
12.2 tok/s | 122.3K 76 | 2 | — | 9.4 |
|
112
Grok 4.7
|
74.4 tok/s | 117.3K 26 | 1 | — | — |
|
113
Llama 3.1 405B
|
26.4 tok/s | 108.3K 27 | 2 | — | — |
|
114
OpenAI o4-mini
|
51.3 tok/s | 13.8K 69 | 2 | — | 26.1 |
|
115
GPT-3.5 Turbo
|
22.5 tok/s | 10.2K 66 | 1 | — | 3.2 |
|
116
Gemma 4 26B
|
8.7 tok/s | 4.0K 23 | 0 | — | 16.7 |
|
117
DeepSeek R1
|
64.1 tok/s | 1.4K 3 | 1 | — | 20.4 |
|
118
DeepSeek V3
|
51.5 tok/s | 1.1K 11 | 1 | — | 14.2 |
|
119
Claude 3.5 Sonnet
|
39.3 tok/s | 319 92 | 2 | — | 9.8 |
|
120
Grok-3 (Beta)
|
21.9 tok/s | 142 5 | 3 | — | — |
|
121
Llama 3.3 70B
|
30.4 tok/s | 107 31 | 1 | — | — |
|
122
GPT-4 Turbo
|
9.2 tok/s | 62 40 | 10 | — | 7.7 |
How to read these statistics
Real usageUsage combines prompt and completion tokens generated by CodingFleet users during the selected period.
Credits and availabilityCredits are charged per request, pro-rata to the tokens used. Models tagged Unlimited cost no credits on the Unlimited plan and above (Unlimited, Elite, Ultimate, Ultimate Max); every other model is charged the credits shown, on every plan. Legacy models are no longer selectable.
IntelligenceA composite score out of 100 covering reasoning, coding and knowledge — the same number the Auto model router weighs against price when it picks a model for you.
