PipelineScore
User Leaderboard

Every run. Every user.

140 submissions from 50 users. Sort, filter, find someone who ran your model.

Want on the board? Run the CLI.
Filter
Allalibabacoheredatabricksdeepseekmetamicrosoftmistralmoonshotnouszhipu|TRUNKMAINLINEFEEDERTAPDRIP|Lab-verified onlyClear filters
#User ↕Model ↕Score ↓Tier ↕
1labLabDeepSeek R1 671B-A37B89.5MAINLINE
2cpu-only-prophetQwen 3 235B-A22B MoE88.8MAINLINE
3flash-attn-fanDeepSeek V3 671B-A37B88.1MAINLINE
4labLabDeepSeek Coder V2 236B87.3MAINLINE
5rig-testerDeepSeek R1 671B-A37B86.6MAINLINE
6mac-stackDeepSeek R1 671B-A37B86.0MAINLINE
7quant-or-dieDeepSeek Coder V2 236B85.3MAINLINE
8gpu-shopperQwen 2.5 72B Instruct85.1MAINLINE
9labLabQwen 3 72B Instruct85.0MAINLINE
10quant-or-dieHermes 3 Llama 3.1 405B84.9MAINLINE
11local-llama-fanHermes 3 Llama 3.1 405B84.8MAINLINE
12cpu-only-prophetQwen 2.5 72B Instruct84.8MAINLINE
13inference-monkDeepSeek Coder V2 236B84.7MAINLINE
14bench-ratDeepSeek Coder V2 236B84.5MAINLINE
15labLabQwen 3 235B-A22B MoE84.0MAINLINE
16deepseek-winsDeepSeek V484.0MAINLINE
17inference-monkDeepSeek R1 Distill Qwen 32B84.0MAINLINE
18sram-hoarderQwen 3 72B Instruct83.8MAINLINE
19labLabDeepSeek V483.7MAINLINE
20reason-firstDeepSeek V483.4MAINLINE
21rust_witchDeepSeek V483.3MAINLINE
22tensor-tomasLlama 3.3 70B Instruct83.3MAINLINE
23tensor-tomasDeepSeek V3 671B-A37B83.2MAINLINE
24labLabHermes 3 Llama 3.1 405B83.1MAINLINE
25judgementalDeepSeek V483.1MAINLINE
26local-llama-fanDeepSeek R1 671B-A37B83.0MAINLINE
27inference-monkQwen 3 235B-A22B MoE83.0MAINLINE
28qwen-stanDeepSeek V482.7MAINLINE
29cohere-cultDeepSeek V482.5MAINLINE
30labLabQwen 3 32B Instruct82.4MAINLINE
31apple-siliconQwen 3 235B-A22B MoE82.3MAINLINE
32rig-testerDeepSeek R1 Distill Qwen 32B82.3MAINLINE
33flash-attn-fanQwen 3 32B Instruct82.3MAINLINE
34edge_runnerDeepSeek V482.2MAINLINE
35labLabCommand A82.0MAINLINE
36quant-or-dieLlama 4 70B Instruct81.9MAINLINE
37open-weightsDeepSeek V481.9MAINLINE
38judgementalDeepSeek V481.9MAINLINE
39mac-stackQwen 2.5 Coder 32B81.8MAINLINE
40gpt-cynicDeepSeek V481.7MAINLINE
41qwen-stanQwen 3.6 72B81.6MAINLINE
42flash-attn-fanHermes 3 Llama 3.1 405B81.5MAINLINE
43open-weightsDeepSeek V481.5MAINLINE
44rocm-riderQwen 3 72B Instruct81.5MAINLINE
45gpt-cynicQwen 3.6 72B81.3MAINLINE
46tokenwranglerQwen 3.6 72B81.2MAINLINE
47labLabLlama 4 70B Instruct81.2MAINLINE
48labLabMixtral 8x22B Instruct81.2MAINLINE
49labLabDeepSeek R1 Distill Qwen 32B81.2MAINLINE
50cli-nativeQwen 3.6 72B81.1MAINLINE
51mac-stackQwen 2.5 32B Instruct81.0MAINLINE
52midnight-bencherLlama 4 70B Instruct81.0MAINLINE
53labLabDeepSeek V3 671B-A37B80.8MAINLINE
54cli-nativeQwen 3.6 72B80.8MAINLINE
55tps-counterQwen 3 72B Instruct80.8MAINLINE
56sram-hoarderQwen 2.5 72B Instruct80.6MAINLINE
57vram-monkDBRX Instruct 132B-MoE80.5MAINLINE
58judgementalQwen 3.6 72B80.5MAINLINE
59tensor-tomasCommand A80.3MAINLINE
60midnight-bencherQwen 2.5 VL 72B80.0MAINLINE
61cli-nativeMistral Large 280.0MAINLINE
62gguf-pilgrimQwen 2.5 Coder 32B80.0MAINLINE
63labLabWizardLM 2 8x22B80.0MAINLINE
64gpt-cynicMistral Large 280.0MAINLINE
65silicon-djinnMistral Large 279.9MAINLINE
66labLabQwen 2.5 72B Instruct79.9MAINLINE
67quant-or-dieQwen 3 14B Instruct79.9MAINLINE
68reason-firstMistral Large 279.8MAINLINE
69labLabKimi K2 Instruct79.8MAINLINE
70rig-testerHermes 3 Llama 3.1 70B79.7MAINLINE
71mistral-maxiQwen 3.6 72B79.6MAINLINE
72gguf-pilgrimDeepSeek V3 671B-A37B79.6MAINLINE
73benchmark-ratQwen 3.6 72B79.5MAINLINE
74gpu-shopperWizardLM 2 8x22B79.5MAINLINE
75labLabQwen 3.6 72B79.4MAINLINE
76tool-callerQwen 3.6 72B79.3MAINLINE
77flash-attn-fanDevstral Small 24B79.3MAINLINE
78one-shot-onlyQwen 3.6 72B79.1MAINLINE
79bench-ratLlama 3.1 70B Instruct79.1MAINLINE
80silicon-djinnQwen 3.6 72B79.1MAINLINE
81quant-or-dieQwen 3 14B Instruct79.0MAINLINE
82gemini-fanMistral Large 279.0MAINLINE
83tensor-tomasDBRX Instruct 132B-MoE78.8MAINLINE
84labLabQwen 2.5 VL 72B78.8MAINLINE
85apple-siliconLlama 3.3 70B Instruct78.8MAINLINE
86temp-zeroMistral Large 278.7MAINLINE
87labLabLlama 3.3 70B Instruct78.7MAINLINE
88labLabDeepSeek V2.578.6MAINLINE
89localmodelsMistral Large 278.6MAINLINE
90sram-hoarderQwen 2.5 Coder 32B78.5MAINLINE
91labLabMistral Large 278.5MAINLINE
92labLabQwen 2.5 Coder 32B78.5MAINLINE
93tensor-tomasDeepSeek R1 Distill Qwen 32B78.5MAINLINE
94sram-hoarderLlama 4 70B Instruct78.3MAINLINE
95flash-attn-fanMixtral 8x22B Instruct78.3MAINLINE
96gguf-pilgrimHermes 3 Llama 3.1 70B78.1MAINLINE
97labLabHermes 3 Llama 3.1 70B78.0MAINLINE
98labLabGLM 4 Plus77.9MAINLINE
99labLabCodestral 22B77.9MAINLINE
100mlx-mikeWizardLM 2 8x22B77.8MAINLINE
1–100 of 140
Next →

Top reviewers

By best PipelineScore submitted. Click a user to open their dashboard.

#1
lab
96 runs
89.5
#2
cpu-only-prophet
17 runs
88.8
#3
flash-attn-fan
17 runs
88.1
#4
rig-tester
15 runs
86.6
#5
mac-stack
21 runs
86.0
#6
quant-or-die
18 runs
85.3
#7
gpu-shopper
12 runs
85.1
#8
local-llama-fan
15 runs
84.8
#9
inference-monk
15 runs
84.7
#10
bench-rat
12 runs
84.5
#11
deepseek-wins
1 run
84.0
#12
sram-hoarder
11 runs
83.8
← Switch to model leaderboard