Generation

Subcategory of Language. 91 models scored.

Model Leaderboard

All models ranked by their Generation subcategory score.

# Model Generation Language Overall
1 Claude Sonnet 4.6 100.00% 100.00% 91.15%
2 o4 Mini 100.00% 80.00% 88.35%
3 Gemini 3 Flash (Preview) 100.00% 95.00% 85.35%
4 DeepSeek-V2 Chat 100.00% 100.00% 84.83%
5 Stealth: Aurora Alpha 100.00% 92.50% 83.79%
6 GPT-4o, Aug. 6th (temp=0) 100.00% 75.00% 82.45%
7 GPT-4o Mini (temp=1) 100.00% 77.50% 79.08%
8 GPT-4o Mini (temp=0) 100.00% 75.00% 78.29%
9 Z.AI GLM 4.5 99.67% 97.33% 86.27%
10 o4 Mini High 99.51% 79.76% 90.29%
11 GPT-4o, Aug. 6th (temp=1) 99.43% 82.21% 82.62%
12 GPT-5 Nano 99.35% 77.18% 82.60%
13 Claude Opus 4.5 99.31% 99.66% 89.69%
14 GPT-4.1 Mini 99.27% 89.64% 83.20%
15 Hermes 3 405B 99.14% 99.57% 82.86%
16 DeepSeek V3.1 98.74% 96.87% 82.39%
17 GPT-5 98.01% 91.50% 91.93%
18 GPT-5 Mini 97.98% 96.49% 92.62%
19 GPT-4.1 97.82% 93.91% 88.68%
20 Gemini 2.5 Flash 97.46% 86.23% 80.60%
21 GPT-4o, May 13th (temp=0) 97.44% 98.72% 85.36%
22 GPT-5.2 97.39% 91.19% 90.26%
23 GPT-5.1 97.28% 93.64% 92.54%
24 Gemini 2.5 Flash (Reasoning) 97.13% 86.06% 86.51%
25 Minimax M2.5 97.10% 96.05% 88.71%
26 Claude 3.5 Sonnet 96.24% 85.62% 84.24%
27 Grok 4 96.22% 90.61% 88.12%
28 Claude 3.7 Sonnet 95.89% 92.95% 83.39%
29 Claude Sonnet 4.6 (Reasoning) 95.15% 97.58% 93.66%
30 Gemini 2.5 Pro 95.14% 92.57% 88.53%
31 Gemini 3 Flash (Preview, Reasoning) 94.87% 94.93% 90.50%
32 Gemini 3.1 Pro (Preview) 94.80% 94.90% 94.37%
33 MoonshotAI: Kimi K2.5 94.19% 97.10% 91.04%
34 Z.AI GLM 4.6 93.20% 96.60% 89.11%
35 GPT-4.1 Nano 92.91% 78.95% 71.94%
36 Claude Sonnet 4 92.61% 91.31% 88.72%
37 Aion 2.0 92.34% 96.17% 89.21%
38 Claude Opus 4.6 92.27% 96.13% 92.35%
39 Claude Opus 4.6 (Reasoning) 92.23% 96.12% 95.02%
40 ByteDance Seed 1.6 91.26% 95.63% 90.70%
41 Qwen 3.5 27B 91.05% 95.52% 90.85%
42 Z.AI GLM 4.7 90.93% 85.46% 88.69%
43 Qwen 3.5 Plus (2026-02-15) 90.21% 95.10% 85.96%
44 Gemma 3 12B 90.19% 80.10% 78.41%
45 GPT-4o, May 13th (temp=1) 90.04% 92.52% 83.80%
46 Qwen 3.5 122B 90.03% 95.01% 91.53%
47 Qwen 3.5 397B A17B 90.03% 95.01% 91.73%
48 DeepSeek V3.2 90.03% 85.01% 82.25%
49 Claude Sonnet 4.5 89.77% 92.39% 88.03%
50 Gemma 3 27B 89.42% 77.21% 77.85%
51 Z.AI GLM 5 89.12% 92.06% 91.23%
52 Qwen 3.5 Flash 88.89% 91.94% 86.38%
53 Claude Haiku 4.5 88.67% 91.84% 85.14%
54 Hermes 3 70B 88.33% 81.66% 72.57%
55 Grok 4.1 Fast 87.51% 88.76% 89.55%
56 Mistral Large 87.27% 88.64% 80.15%
57 Claude Opus 4 86.02% 93.01% 87.69%
58 Gemini 2.5 Flash Lite 85.50% 82.75% 81.08%
59 Llama 3.1 70B 85.36% 80.18% 78.40%
60 Gemini 3 Pro (Preview) 84.29% 89.64% 88.79%
61 Grok 4 Fast 84.22% 84.61% 86.15%
62 Mistral Large 3 84.05% 92.02% 85.43%
63 Qwen 3.5 35B 83.90% 91.95% 88.00%
64 Gemini 2.5 Flash Lite (Reasoning) 83.71% 74.36% 85.75%
65 ByteDance Seed 1.6 Flash 82.45% 61.23% 73.27%
66 Arcee AI: Trinity Mini 81.18% 70.59% 70.90%
67 Claude 3 Haiku 80.53% 72.76% 71.19%
68 Z.AI GLM 4.7 Flash 80.35% 87.67% 84.82%
69 Arcee AI: Trinity Large (Preview) 76.76% 78.38% 73.33%
70 DeepSeek V3 (2024-12-26) 75.76% 87.88% 83.68%
71 Gemma 3 4B 74.56% 72.28% 68.57%
72 Claude 3.5 Haiku 74.25% 82.12% 83.73%
73 Cohere Command R+ (Aug. 2024) 73.15% 66.58% 69.03%
74 Llama 3.1 8B 73.13% 64.06% 63.37%
75 DeepSeek V3 (2025-03-24) 72.83% 86.42% 81.99%
76 Mistral Small 3.2 24B 70.53% 72.77% 78.60%
77 Mistral Large 2 70.43% 85.22% 82.41%
78 Qwen 2.5 72B 67.91% 68.95% 75.46%
79 Mistral NeMO 66.60% 80.80% 65.04%
80 LFM2 24B 64.28% 64.64% 58.77%
81 WizardLM 2 8x22b 61.10% 78.05% 71.07%
82 Ministral 3B 59.49% 42.25% 61.29%
83 Ministral 8B 52.81% 53.91% 64.87%
84 Rocinante 12B 51.90% 63.45% 54.55%
85 Mistral Medium 3.1 49.00% 49.50% 77.83%
86 Ministral 3 8B 47.92% 48.96% 71.76%
87 Writer: Palmyra X5 43.16% 56.58% 79.57%
88 Ministral 3 3B 36.20% 68.10% 67.22%
89 Mistral Small Creative 33.69% 41.85% 73.27%
90 Llama 3.1 Nemotron 70B 33.61% 46.80% 74.70%
91 Ministral 3 14B 10.00% 30.00% 72.54%