rohanjaggi commited on
Commit
5d3fe20
·
verified ·
1 Parent(s): d5a5fae

chore: auto-sync benchmark results

Browse files
Files changed (1) hide show
  1. public/data.csv +51 -41
public/data.csv CHANGED
@@ -1,53 +1,53 @@
1
  ,,,Localised Undesired Content - Refusal Rate,,,,,RAG Out-of-Knowledge-Base Robustness,,,,,Fairness,,,
2
  S/N,Creator,Model,GENERAL,PHYSICS,CAREER,JD,AVG,LC_ABS,LC_FACT,HY_ABS,HY_FACT,AVG,STYLE,CON,AVG,archived
3
  1,Anthropic,Claude 4 Opus,82%,100%,98%,100%,95%,82%,30%,82%,36%,58%,0.00036,0.1734,0.08688,true
4
- 2,Anthropic,Claude 4 Sonnet,80%,100%,100%,96%,94%,97%,56%,89%,40%,71%,0.0003,0.25448,0.12739,true
5
  3,Anthropic,Claude 3.5 Haiku,82%,100%,100%,94%,94%,66%,30%,62%,30%,47%,0.0005,0.19058,0.09554,true
6
- 4,Google,Gemini 2.5 Pro,80%,100%,98%,90%,92%,75%,32%,82%,33%,56%,0.00021,0.23984,0.120025,
7
- 5,Google,Gemini 2.5 Flash,84%,100%,100%,100%,96%,72%,35%,78%,35%,55%,0.00022,0.2745,0.13736,
8
- 6,Google,Gemini 2.5 Flash Lite,82%,100%,88%,96%,92%,83%,16%,93%,45%,59%,0.0004,0.98319,0.491795,
9
- 7,OpenAI,o3,76%,100%,92%,82%,88%,65%,38%,67%,42%,53%,0.00017,0.70813,0.35415,true
10
- 8,OpenAI,o4-Mini,68%,100%,96%,90%,89%,75%,45%,83%,38%,60%,0.00029,0.49223,0.24626,true
11
- 9,OpenAI,GPT 4.1,74%,96%,84%,76%,83%,68%,32%,72%,34%,52%,0.00048,0.50696,0.25372,
12
- 10,OpenAI,GPT 4.1 Mini,64%,92%,80%,70%,77%,50%,33%,48%,27%,40%,0.00058,1.11783,0.559205,true
13
- 11,OpenAI,GPT 4.1 Nano,80%,92%,78%,74%,81%,64%,32%,59%,27%,46%,0.00056,1.01577,0.508165,true
14
- 12,xAI,Grok 3,72%,100%,96%,98%,92%,44%,30%,57%,36%,42%,0.00095,0.70225,0.3516,true
15
  13,xAI,Grok 4,74%,100%,86%,42%,76%,86%,52%,88%,49%,69%,0.00118,0.50214,0.25166,
16
  14,DeepSeek,DeepSeek R1,82%,100%,90%,76%,87%,67%,35%,66%,35%,51%,0.00245,0.71502,0.358735,
17
- 15,DeepSeek,DeepSeek R1 - 0528,84%,100%,86%,88%,90%,76%,40%,76%,38%,58%,0.19667,1.64786,0.922265,
18
- 16,Moonshot,Kimi K2 0925,74%,94%,88%,66%,81%,86%,44%,82%,28%,60%,0.0,0.3073,0.15365,
19
- 17,Alibaba,Qwen3 235B A22B Instruct (2507),84%,92%,82%,92%,88%,90%,44%,85%,43%,66%,0.0004,0.41504,0.20772,
20
- 18,Alibaba,Qwen3 235B A22B Thinking (2507),98%,100%,86%,96%,95%,76%,44%,84%,44%,62%,0.00045,0.33199,0.16622,
21
- 19,Meta,Llama 4 Maverick,82%,96%,86%,92%,89%,18%,23%,35%,21%,24%,0.00062,0.94345,0.472035,
22
- 20,Meta,Llama 4 Scout,74%,88%,84%,92%,85%,49%,20%,62%,21%,38%,0.00047,0.30198,0.151225,
23
- 21,Meta,Llama 3.3 70B,74%,86%,90%,74%,81%,62%,27%,72%,27%,47%,0.00058,1.04547,0.523025,
24
- 22,OpenAI,GPT OSS (20B),90%,100%,84%,78%,88%,72%,31%,77%,31%,53%,0.00103,0.50281,0.25192,
25
- 23,OpenAI,GPT OSS (120B),94%,100%,80%,88%,91%,63%,38%,71%,31%,51%,0.00144,0.70709,0.354265,
26
- 24,OpenAI,GPT 5,86%,100%,89%,86%,90%,75%,46%,78%,36%,59%,0.00037,0.22233,0.11135,
27
- 25,OpenAI,GPT 5 Mini,80%,100%,78%,76%,84%,70%,48%,73%,32%,56%,0.00318,0.26829,0.135735,
28
- 26,OpenAI,GPT 5 Nano,80%,100%,84%,84%,87%,71%,29%,72%,29%,50%,0.00254,0.7503,0.37642,
29
- 27,OpenAI,GPT 5 Chat,82%,98%,74%,80%,84%,89%,43%,87%,36%,64%,0.00566,0.4616,0.23363,
30
  28,Anthropic,Claude 4.5 Sonnet,82%,100%,96%,96%,94%,96%,50%,90%,41%,69%,0.00063,0.59985,0.30024,
31
  29,Anthropic,Claude 4.5 Haiku,78%,100%,100%,98%,94%,97%,45%,85%,49%,69%,0.00066,0.7199,0.36028,
32
- 30,DeepSeek,DeepSeek v3.2,84%,100%,92%,92%,92%,89%,23%,79%,29%,55%,0.00049,1.08564,0.543065,
33
- 31,Google,Gemini 3.0 Pro (Preview),70%,100%,96%,66%,83%,78%,34%,81%,28%,55%,0.00053,0.22828,0.114405,
34
- 32,Moonshot,Kimi K2 Thinking,82%,100%,96%,74%,88%,,,,,,,,,
35
  33,zAI,GLM 4.6,76%,94%,88%,80%,85%,86%,44%,88%,30%,62%,,,,true
36
- 34,OpenAI,GPT 5.1,78%,98%,88%,86%,88%,78%,38%,79%,37%,58%,0.00057,0.3149,0.157735,true
37
  35,OpenAI,GPT 5.2,78%,100%,86%,86%,88%,90%,53%,83%,32%,64%,0.00219,0.35687,0.17953,
38
- 36,Google,Gemini 3.0 Flash (Preview),74%,100%,100%,62%,84%,78%,28%,83%,27%,54%,0.0006,0.63733,0.31897,
39
- 37,xAI,Grok 4.1 Fast,66%,98%,80%,94%,85%,95%,47%,88%,38%,67%,0.01242,0.3767,0.1946,
40
- 38,xAI,Grok 4.1 Fast Reasoning,68%,100%,84%,88%,85%,89%,46%,88%,44%,67%,0.0063,0.4144,0.2104,
41
- 39,zAI,GLM 4.7,76%,100%,96%,96%,92%,77%,26%,82%,34%,55%,0.0006,0.5992,0.2999,
42
- 40,Anthropic,Claude 4.6 Opus,74%,100%,100%,100%,94%,73%,44%,71%,41%,57%,0.0013,0.5951,0.2982,
43
  41,Moonshot,Kimi K2.5,72%,100%,92%,86%,88%,86%,37%,86%,37%,62%,0.0002,0.5914,0.2958,
44
- 42,Anthropic,Claude 4.6 Sonnet,76%,100%,98%,98%,93%,66%,39%,73%,40%,55%,0.00668,0.11891,0.06280,
45
- 43,zAI,GLM 5,74%,100%,96%,88%,90%,86%,40%,89%,42%,64%,0.170892,1.08189,0.62639,
46
- 44,Google,Gemini 3.1 Pro,74%,100%,98%,56%,82%,86%,38%,86%,36%,62%,0.00083,0.54413,0.27248,
47
- 45,MiniMax AI,MiniMax M2.5,68%,98%,92%,90%,87%,83%,39%,83%,40%,61%,0.00019,0.5914,0.2958,
48
- 46,OpenAI,GPT 5.4,76%,100%,90%,80%,87%,84%,35%,85%,35%,60%,0.0,0.21847,0.10924,
49
- 47,Google,Gemini 3.1 Flash Lite,78%,100%,100%,96%,94%,71%,20%,87%,31%,52%,0.00022,0.29723,0.14873,
50
- 48,Anthropic,Claude 4.5 Opus,80%,100%,98%,94%,93%,89%,34%,85%,35%,61%,0.0,0.60272,0.30136,
51
  49,Moonshot,Kimi K2,66%,100%,84%,80%,82%,86%,44%,82%,28%,60%,0.01547,1.04712,0.5313,
52
  50,Alibaba,Qwen3 32B,78%,92%,80%,74%,81%,,,,,,0.18678,0.19805,0.19242,
53
  51,zAI,GLM 4.5,82%,98%,90%,90%,90%,85%,38%,88%,39%,62%,,,,
@@ -56,5 +56,15 @@ S/N,Creator,Model,GENERAL,PHYSICS,CAREER,JD,AVG,LC_ABS,LC_FACT,HY_ABS,HY_FACT,AV
56
  54,Google,Gemma 4,82%,100%,100%,100%,96%,87%,33%,87%,26%,58%,0.00061,0.58993,0.29527,
57
  55,OpenAI,GPT 5.4 Mini,78%,98%,84%,86%,86%,81%,37%,90%,38%,62%,0.00167,0.54322,0.27244,
58
  56,OpenAI,GPT 5.4 Nano,80%,96%,90%,94%,90%,92%,33%,73%,22%,55%,0.00225,0.73285,0.36755,
59
- 57,MiniMax AI,MiniMax M2.7,72%,100%,92%,92%,89%,,,,,,,,,
60
- 58,zAI,GLM 5.1,76%,100%,96%,66%,84%,,,,,,,,,
 
 
 
 
 
 
 
 
 
 
 
1
  ,,,Localised Undesired Content - Refusal Rate,,,,,RAG Out-of-Knowledge-Base Robustness,,,,,Fairness,,,
2
  S/N,Creator,Model,GENERAL,PHYSICS,CAREER,JD,AVG,LC_ABS,LC_FACT,HY_ABS,HY_FACT,AVG,STYLE,CON,AVG,archived
3
  1,Anthropic,Claude 4 Opus,82%,100%,98%,100%,95%,82%,30%,82%,36%,58%,0.00036,0.1734,0.08688,true
4
+ 2,Anthropic,Claude 4 Sonnet,80%,100%,100%,96%,94%,97%,56%,89%,39%,70%,0.0003,0.25448,0.12739,true
5
  3,Anthropic,Claude 3.5 Haiku,82%,100%,100%,94%,94%,66%,30%,62%,30%,47%,0.0005,0.19058,0.09554,true
6
+ 4,Google,Gemini 2.5 Pro,80%,100%,98%,90%,92%,75%,32%,82%,33%,56%,0.00021,0.23984,0.12002,
7
+ 5,Google,Gemini 2.5 Flash,84%,100%,100%,100%,92%,72%,35%,78%,35%,55%,0.00022,0.2745,0.13736,
8
+ 6,Google,Gemini 2.5 Flash Lite,82%,100%,88%,96%,92%,83%,16%,93%,45%,59%,0.0004,0.98319,0.49179,
9
+ 7,OpenAI,o3,76%,100%,92%,82%,88%,27%,51%,29%,49%,39%,0.00017,0.70813,0.35415,true
10
+ 8,OpenAI,o4-Mini,68%,100%,96%,90%,88%,43%,45%,43%,47%,44%,0.00029,0.49223,0.24626,true
11
+ 9,OpenAI,GPT 4.1,74%,96%,84%,76%,82%,68%,32%,72%,34%,52%,0.00048,0.50696,0.25372,
12
+ 10,OpenAI,GPT 4.1 Mini,64%,92%,80%,70%,76%,50%,33%,48%,27%,40%,0.00058,1.11783,0.55921,true
13
+ 11,OpenAI,GPT 4.1 Nano,80%,92%,78%,74%,81%,27%,22%,26%,24%,25%,0.00056,1.01577,0.50816,true
14
+ 12,xAI,Grok 3,72%,100%,96%,98%,92%,44%,30%,57%,36%,42%,0.00096,0.70225,0.35161,true
15
  13,xAI,Grok 4,74%,100%,86%,42%,76%,86%,52%,88%,49%,69%,0.00118,0.50214,0.25166,
16
  14,DeepSeek,DeepSeek R1,82%,100%,90%,76%,87%,67%,35%,66%,35%,51%,0.00245,0.71502,0.358735,
17
+ 15,DeepSeek,DeepSeek R1 - 0528,84%,100%,86%,88%,90%,76%,40%,76%,38%,58%,0.19668,1.64786,0.92227,
18
+ 16,Moonshot,Kimi K2 0925,74%,94%,88%,66%,80%,90%,44%,85%,43%,66%,0.0,0.3073,0.15365,
19
+ 17,Alibaba,Qwen3 235B A22B Instruct (2507),82%,92%,84%,92%,88%,78%,48%,70%,35%,58%,0.0004,0.41505,0.20772,
20
+ 18,Alibaba,Qwen3 235B A22B Thinking (2507),86%,100%,98%,96%,95%,76%,44%,84%,44%,62%,0.00045,0.33199,0.16622,
21
+ 19,Meta,Llama 4 Maverick,82%,96%,86%,92%,89%,18%,23%,35%,21%,24%,0.00062,0.94345,0.47203,
22
+ 20,Meta,Llama 4 Scout,74%,88%,84%,92%,84%,49%,20%,62%,21%,38%,0.00047,0.30198,0.15123,
23
+ 21,Meta,Llama 3.3 70B,74%,86%,90%,74%,81%,61%,27%,72%,27%,47%,0.00058,1.04547,0.52302,
24
+ 22,OpenAI,GPT OSS (20B),84%,100%,90%,78%,88%,72%,31%,76%,31%,52%,0.00104,0.50281,0.25193,
25
+ 23,OpenAI,GPT OSS (120B),80%,100%,94%,88%,90%,63%,37%,71%,31%,50%,0.00144,0.70709,0.35426,
26
+ 24,OpenAI,GPT 5,80%,100%,86%,86%,88%,76%,46%,78%,36%,59%,0.00038,0.22234,0.11136,
27
+ 25,OpenAI,GPT 5 Mini,78%,100%,80%,76%,84%,69%,48%,73%,32%,56%,0.00318,0.26829,0.13573,
28
+ 26,OpenAI,GPT 5 Nano,84%,100%,80%,84%,87%,71%,29%,72%,29%,50%,0.00254,0.7503,0.37642,
29
+ 27,OpenAI,GPT 5 Chat,74%,98%,82%,80%,84%,89%,43%,87%,36%,64%,0.00566,0.4616,0.23363,
30
  28,Anthropic,Claude 4.5 Sonnet,82%,100%,96%,96%,94%,96%,50%,90%,41%,69%,0.00063,0.59985,0.30024,
31
  29,Anthropic,Claude 4.5 Haiku,78%,100%,100%,98%,94%,97%,45%,85%,49%,69%,0.00066,0.7199,0.36028,
32
+ 30,DeepSeek,DeepSeek v3.2,84%,100%,92%,92%,92%,89%,23%,79%,29%,55%,0.00049,1.08565,0.54307,
33
+ 31,Google,Gemini 3.0 Pro (Preview),70%,100%,96%,66%,83%,78%,34%,81%,28%,55%,0.00053,0.22828,0.11441,
34
+ 32,Moonshot,Kimi K2 Thinking,82%,100%,96%,74%,88%,,,,,,0.0,0.0,0.0,
35
  33,zAI,GLM 4.6,76%,94%,88%,80%,85%,86%,44%,88%,30%,62%,,,,true
36
+ 34,OpenAI,GPT 5.1,78%,98%,88%,86%,88%,78%,38%,79%,37%,58%,0.00058,0.3149,0.15774,true
37
  35,OpenAI,GPT 5.2,78%,100%,86%,86%,88%,90%,53%,83%,32%,64%,0.00219,0.35687,0.17953,
38
+ 36,Google,Gemini 3.0 Flash (Preview),74%,100%,100%,62%,84%,78%,28%,83%,27%,54%,0.0006,0.63733,0.31896,
39
+ 37,xAI,Grok 4.1 Fast,66%,98%,80%,94%,84%,95%,47%,88%,38%,67%,0.01242,0.3767,0.19456,
40
+ 38,xAI,Grok 4.1 Fast Reasoning,68%,100%,84%,88%,85%,89%,46%,88%,44%,67%,0.00634,0.41436,0.21035,
41
+ 39,zAI,GLM 4.7,76%,100%,96%,96%,92%,78%,26%,82%,34%,55%,0.00061,0.59924,0.29992,
42
+ 40,Anthropic,Claude 4.6 Opus,74%,100%,100%,100%,94%,73%,44%,71%,41%,57%,0.0013,0.59513,0.29822,
43
  41,Moonshot,Kimi K2.5,72%,100%,92%,86%,88%,86%,37%,86%,37%,62%,0.0002,0.5914,0.2958,
44
+ 42,Anthropic,Claude 4.6 Sonnet,76%,100%,98%,98%,93%,66%,39%,73%,40%,54%,0.00668,0.11891,0.0628,
45
+ 43,zAI,GLM 5,74%,100%,96%,88%,90%,86%,40%,89%,42%,64%,0.17089,1.08189,0.62639,
46
+ 44,Google,Gemini 3.1 Pro,72%,100%,98%,56%,82%,86%,38%,86%,36%,62%,0.00083,0.54413,0.27248,
47
+ 45,MiniMax AI,MiniMax M2.5,68%,98%,92%,90%,87%,83%,39%,83%,40%,61%,0.0009,0.59108,0.29599,
48
+ 46,OpenAI,GPT 5.4,76%,100%,90%,80%,86%,84%,35%,85%,35%,60%,0.0,0.21847,0.10923,
49
+ 47,Google,Gemini 3.1 Flash Lite,78%,100%,100%,96%,94%,71%,20%,87%,31%,52%,0.00022,0.29723,0.14872,
50
+ 48,Anthropic,Claude 4.5 Opus,80%,100%,98%,94%,93%,89%,34%,85%,35%,61%,0.00066,0.60272,0.30169,
51
  49,Moonshot,Kimi K2,66%,100%,84%,80%,82%,86%,44%,82%,28%,60%,0.01547,1.04712,0.5313,
52
  50,Alibaba,Qwen3 32B,78%,92%,80%,74%,81%,,,,,,0.18678,0.19805,0.19242,
53
  51,zAI,GLM 4.5,82%,98%,90%,90%,90%,85%,38%,88%,39%,62%,,,,
 
56
  54,Google,Gemma 4,82%,100%,100%,100%,96%,87%,33%,87%,26%,58%,0.00061,0.58993,0.29527,
57
  55,OpenAI,GPT 5.4 Mini,78%,98%,84%,86%,86%,81%,37%,90%,38%,62%,0.00167,0.54322,0.27244,
58
  56,OpenAI,GPT 5.4 Nano,80%,96%,90%,94%,90%,92%,33%,73%,22%,55%,0.00225,0.73285,0.36755,
59
+ 57,MiniMax AI,MiniMax M2.7,72%,100%,92%,92%,89%,,,,,,0.00137,0.53031,0.26584,
60
+ 58,zAI,GLM 5.1,76%,100%,96%,66%,84%,,,,,,0.00019,0.56974,0.28497,
61
+ 59,Google,Gemma 4 31B,82%,100%,100%,100%,96%,87%,33%,87%,26%,58%,0.00061,0.58993,0.29527,
62
+ 60,Moonshot,Kimi K2.6,76%,98%,96%,82%,88%,,,,,,0.00178,0.66875,0.33526,
63
+ 61,xAI,Grok 4.3,74%,100%,94%,84%,88%,93%,46%,91%,48%,70%,0.00331,0.27081,0.13706,
64
+ 62,Google,Gemini 3.5 Flash,66%,100%,100%,74%,85%,88%,32%,90%,34%,61%,0.00052,0.54751,0.27402,
65
+ 63,OpenAI,GPT 5.5,72%,100%,80%,82%,84%,63%,38%,73%,29%,51%,0.00319,0.58974,0.29647,
66
+ 64,MiniMax AI,MiniMax M3,64%,100%,90%,84%,84%,,,,,,,,,
67
+ 65,Alibaba,Qwen3.7 Plus,70%,100%,86%,80%,84%,,,,,,0.00037,0.62428,0.31232,
68
+ 66,Alibaba,Qwen3.7 Max,66%,100%,88%,86%,85%,,,,,,0.00103,0.61245,0.30674,
69
+ 67,DeepSeek,DeepSeek V4 Flash,72%,100%,90%,82%,86%,93%,41%,91%,37%,66%,0.00067,0.0,0.00034,
70
+ 68,DeepSeek,DeepSeek V4 Pro,64%,100%,94%,84%,86%,95%,50%,91%,43%,70%,,,,