Model
Fit
Speed
Memory
Scores / Run
Runs comfortably
Ornith-1.5-35B-A3B
Q8_0
37.8GB
262k ctx
Runs comfortably
163
tok/s
ⓘ estimated
38
/ 48GB
91
/
89
/
86
Runs comfortably
163 tok/s
ⓘ estimated
38/48GB
Ornith-1.5-35B-A3B
Q6_K
29.2GB
262k ctx
Runs comfortably
211
tok/s
ⓘ estimated
29
/ 48GB
91
/
89
/
86
Runs comfortably
211 tok/s
ⓘ estimated
29/48GB
Ornith-1.5-35B-A3B
Q5_K_M
25.3GB
262k ctx
Runs comfortably
244
tok/s
ⓘ estimated
25
/ 48GB
91
/
89
/
86
Runs comfortably
244 tok/s
ⓘ estimated
25/48GB
Ornith-1.5-35B-A3B
Q4_K_M
21.7GB
262k ctx
Runs comfortably
284
tok/s
ⓘ estimated
22
/ 48GB
91
/
89
/
86
Runs comfortably
284 tok/s
ⓘ estimated
22/48GB
Qwen3.6 35B A3B
Q4_K_M
20.0GB
262k ctx
Runs comfortably
300
tok/s
ⓘ estimated
20
/ 48GB
90
/
91
/
87
Runs comfortably
300 tok/s
ⓘ estimated
20/48GB
Runs comfortably
162 tok/s
ⓘ estimated
37/48GB
Pokee-Isaac 28B
Q4_K_M
16.0GB
10000k ctx
Runs comfortably
32
tok/s
ⓘ estimated
16
/ 48GB
90
/
88
/
88
Runs comfortably
32 tok/s
ⓘ estimated
16/48GB
Runs comfortably
18 tok/s
ⓘ estimated
28/48GB
17.6GB
262k ctx
Runs comfortably
29
tok/s
ⓘ estimated
18
/ 48GB
88
/
89
/
85
Runs comfortably
29 tok/s
ⓘ estimated
18/48GB
9.8GB
262k ctx
Runs comfortably
52
tok/s
ⓘ estimated
10
/ 48GB
88
/
89
/
85
Runs comfortably
52 tok/s
ⓘ estimated
10/48GB
6.2GB
262k ctx
Runs comfortably
83
tok/s
ⓘ estimated
6
/ 48GB
88
/
89
/
85
Runs comfortably
83 tok/s
ⓘ estimated
6/48GB
Runs comfortably
16 tok/s
ⓘ estimated
33/48GB
Runs comfortably
29 tok/s
ⓘ estimated
18/48GB
Runs comfortably
28 tok/s
ⓘ estimated
18/48GB
Runs comfortably
136 tok/s
ⓘ estimated
25/48GB
Gemma 4 26B A4B
Q4_K_M
13.0GB
256k ctx
Runs comfortably
262
tok/s
ⓘ estimated
13
/ 48GB
85
/
88
/
84
Runs comfortably
262 tok/s
ⓘ estimated
13/48GB
Runs comfortably
31 tok/s
ⓘ estimated
17/48GB
Runs comfortably
18 tok/s
ⓘ estimated
28/48GB
Runs comfortably
18 tok/s
ⓘ estimated
29/48GB
Runs comfortably
21 tok/s
ⓘ estimated
25/48GB
Runs comfortably
26 tok/s
ⓘ estimated
20/48GB
Runs comfortably
15 tok/s
ⓘ estimated
34/48GB
Runs comfortably
28 tok/s
ⓘ estimated
18/48GB
Runs comfortably
53 tok/s
ⓘ estimated
10/48GB
Runs comfortably
68 tok/s
ⓘ estimated
8/48GB
Runs comfortably
78 tok/s
ⓘ estimated
7/48GB
Runs comfortably
89 tok/s
ⓘ estimated
6/48GB
Nemotron 3.5 Lightning
Q4_K_M
32.0GB
1000k ctx
Runs comfortably
141
tok/s
ⓘ estimated
32
/ 48GB
78
/
76
/
74
Runs comfortably
141 tok/s
ⓘ estimated
32/48GB
Runs comfortably
337 tok/s
ⓘ estimated
16/48GB
Runs comfortably
169 tok/s
ⓘ estimated
31/48GB
Runs comfortably
18 tok/s
ⓘ estimated
29/48GB
Runs comfortably
61 tok/s
ⓘ estimated
9/48GB
Runs comfortably
34 tok/s
ⓘ estimated
15/48GB
21.5GB
128k ctx
Runs comfortably
24
tok/s
ⓘ estimated
22
/ 48GB
72
/
77
/
80
Runs comfortably
24 tok/s
ⓘ estimated
22/48GB
Llama 3.3 70B Instruct
Q3_K_M
30.0GB
128k ctx
Runs comfortably
17
tok/s
ⓘ estimated
30
/ 48GB
72
/
77
/
80
Runs comfortably
17 tok/s
ⓘ estimated
30/48GB
Llama 3.3 70B Instruct
Q4_K_M
42.0GB
128k ctx
Runs comfortably
12
tok/s
ⓘ estimated
42
/ 48GB
72
/
77
/
80
Runs comfortably
12 tok/s
ⓘ estimated
42/48GB
Runs comfortably
34 tok/s
ⓘ estimated
15/48GB
Runs comfortably
57 tok/s
ⓘ estimated
9/48GB
Runs comfortably
40 tok/s
ⓘ estimated
13/48GB
Runs comfortably
72 tok/s
ⓘ estimated
7/48GB
Runs comfortably
18 tok/s
ⓘ estimated
28/48GB
Runs comfortably
30 tok/s
ⓘ estimated
17/48GB
Runs comfortably
61 tok/s
ⓘ estimated
9/48GB
Runs comfortably
99 tok/s
ⓘ estimated
5/48GB
Mixtral 8x7B Instruct
Q3_K_M
19.0GB
32k ctx
Runs comfortably
98
tok/s
ⓘ estimated
19
/ 48GB
62
/
65
/
68
Runs comfortably
98 tok/s
ⓘ estimated
19/48GB
Mixtral 8x7B Instruct
Q4_K_M
26.5GB
32k ctx
Runs comfortably
70
tok/s
ⓘ estimated
27
/ 48GB
62
/
65
/
68
Runs comfortably
70 tok/s
ⓘ estimated
27/48GB
Runs comfortably
69 tok/s
ⓘ estimated
8/48GB
Runs comfortably
38 tok/s
ⓘ estimated
14/48GB
Runs comfortably
39 tok/s
ⓘ estimated
13/48GB
Runs comfortably
71 tok/s
ⓘ estimated
7/48GB
Runs comfortably
198 tok/s
ⓘ estimated
3/48GB
Runs comfortably
206 tok/s
ⓘ estimated
3/48GB
Runs comfortably
126 tok/s
ⓘ estimated
4/48GB
16.0GB
128k ctx
Runs comfortably
32
tok/s
ⓘ estimated
16
/ 48GB
52
/
55
/
62
Runs comfortably
32 tok/s
ⓘ estimated
16/48GB
8.0GB
128k ctx
Runs comfortably
64
tok/s
ⓘ estimated
8
/ 48GB
52
/
55
/
62
Runs comfortably
64 tok/s
ⓘ estimated
8/48GB
Llama 3.1 8B Instruct
Q4_K_M
4.5GB
128k ctx
Runs comfortably
114
tok/s
ⓘ estimated
5
/ 48GB
52
/
55
/
62
Runs comfortably
114 tok/s
ⓘ estimated
5/48GB
Runs comfortably
103 tok/s
ⓘ estimated
5/48GB
Runs comfortably
172 tok/s
ⓘ estimated
3/48GB
3.5GB
128k ctx
Runs comfortably
147
tok/s
ⓘ estimated
4
/ 48GB
35
/
38
/
50
Runs comfortably
147 tok/s
ⓘ estimated
4/48GB
Llama 3.2 3B Instruct
Q4_K_M
2.0GB
128k ctx
Runs comfortably
257
tok/s
ⓘ estimated
2
/ 48GB
35
/
38
/
50
Runs comfortably
257 tok/s
ⓘ estimated
2/48GB
Runs comfortably
29 tok/s
ⓘ estimated
18/48GB
Runs comfortably
16 tok/s
ⓘ estimated
33/48GB
Runs with CPU offload
72.5GB
0k cap
Runs with CPU offload
148
tok/s
ⓘ estimated
73
/ 48GB
92
/
92
/
89
Runs with CPU offload
148 tok/s
ⓘ estimated
73/48GB
Ornith-1.5-35B-A3B
BF16
71.1GB
0k cap
Runs with CPU offload
87
tok/s
ⓘ estimated
71
/ 48GB
91
/
89
/
86
Runs with CPU offload
87 tok/s
ⓘ estimated
71/48GB
Runs with CPU offload
9 tok/s
ⓘ estimated
56/48GB
Qwen3 235B A22B
Q2_K
68.0GB
0k cap
Runs with CPU offload
81
tok/s
ⓘ estimated
68
/ 48GB
88
/
90
/
85
Runs with CPU offload
81 tok/s
ⓘ estimated
68/48GB
Runs with CPU offload
8 tok/s
ⓘ estimated
61/48GB
Gemma 4 26B A4B
BF16
50.0GB
0k cap
Runs with CPU offload
68
tok/s
ⓘ estimated
50
/ 48GB
85
/
88
/
84
Runs with CPU offload
68 tok/s
ⓘ estimated
50/48GB
60.0GB
0k cap
Runs with CPU offload
75
tok/s
ⓘ estimated
60
/ 48GB
78
/
76
/
74
Runs with CPU offload
75 tok/s
ⓘ estimated
60/48GB
58.0GB
0k cap
Runs with CPU offload
78
tok/s
ⓘ estimated
58
/ 48GB
78
/
76
/
74
Runs with CPU offload
78 tok/s
ⓘ estimated
58/48GB
74.0GB
0k cap
Runs with CPU offload
7
tok/s
ⓘ estimated
74
/ 48GB
72
/
77
/
80
Runs with CPU offload
7 tok/s
ⓘ estimated
74/48GB
Coding alternatives
Hosted coding assistants for comparison - per seat
Cursor
Hobby
AI-first code editor. Uses frontier models (GPT-4, Claude). $20/user Business...
$0.0
GitHub Copilot
Individual
AI coding assistant by GitHub. GPT-4o and Claude-based completions. $19/seat ...
$10.0
GitHub Copilot
Business
AI coding assistant by GitHub. GPT-4o and Claude-based completions. $19/seat ...
$19.0
Cursor
Pro
AI-first code editor. Uses frontier models (GPT-4, Claude). $20/user Business...
$20.0
GitHub Copilot
Enterprise
AI coding assistant by GitHub. GPT-4o and Claude-based completions. $19/seat ...
$39.0
Cursor
Business
AI-first code editor. Uses frontier models (GPT-4, Claude). $20/user Business...
$40.0
Per-token monthly figures are estimates at typical usage. Prices verified periodically - always confirm current pricing on the provider's website.
Sign in with GitHub to save this rig and get weekly model updates
Sign in with GitHub