Sentence Counting

Subcategory of Utility. 91 models scored.

Model Leaderboard

All models ranked by their Sentence Counting subcategory score.

# Model Sentence Counting Utility Overall
1 Gemini 3.1 Pro (Preview) 100.00% 99.91% 94.37%
2 GPT-5 100.00% 93.53% 91.93%
3 Qwen 3.5 397B A17B 100.00% 97.50% 91.73%
4 Qwen 3.5 122B 100.00% 96.36% 91.53%
5 Z.AI GLM 5 100.00% 94.11% 91.23%
6 Qwen 3.5 35B 100.00% 96.42% 88.00%
7 Qwen 3.5 Flash 100.00% 96.11% 86.38%
8 Qwen 3.5 Plus (2026-02-15) 100.00% 86.65% 85.96%
9 GPT-5 Mini 100.00% 98.39% 92.62%
10 MoonshotAI: Kimi K2.5 100.00% 96.63% 91.04%
11 ByteDance Seed 1.6 100.00% 90.83% 90.70%
12 Claude Opus 4.6 (Reasoning) 100.00% 98.93% 95.02%
13 Gemini 3 Flash (Preview, Reasoning) 100.00% 97.20% 90.50%
14 Z.AI GLM 4.7 Flash 100.00% 88.98% 84.82%
15 Z.AI GLM 4.7 99.99% 94.31% 88.69%
16 Gemini 3 Flash (Preview) 99.99% 86.39% 85.35%
17 Llama 3.1 Nemotron 70B 99.99% 88.31% 74.70%
18 Gemini 3 Pro (Preview) 99.97% 96.14% 88.79%
19 Mistral Small 3.2 24B 99.97% 73.17% 78.60%
20 Claude Sonnet 4.6 (Reasoning) 99.96% 97.88% 93.66%
21 Mistral Large 3 99.95% 84.91% 85.43%
22 Claude Opus 4.6 99.93% 90.72% 92.35%
23 o4 Mini High 99.90% 98.67% 90.29%
24 GPT-5.1 99.80% 95.33% 92.54%
25 GPT-5.2 99.77% 96.22% 90.26%
26 Gemini 2.5 Pro 99.43% 92.18% 88.53%
27 o4 Mini 98.35% 96.31% 88.35%
28 GPT-5 Nano 97.99% 93.91% 82.60%
29 DeepSeek V3 (2025-03-24) 97.96% 80.62% 81.99%
30 DeepSeek-V2 Chat 97.87% 83.82% 84.83%
31 Z.AI GLM 4.6 97.70% 88.58% 89.11%
32 Stealth: Aurora Alpha 97.37% 92.59% 83.79%
33 Mistral Medium 3.1 97.23% 80.13% 77.83%
34 Mistral Small Creative 96.83% 76.28% 73.27%
35 Gemini 2.5 Flash Lite (Reasoning) 96.54% 89.63% 85.75%
36 Claude Opus 4.5 95.88% 89.84% 89.69%
37 Grok 4.1 Fast 95.61% 84.12% 89.55%
38 Gemma 3 27B 95.44% 76.82% 77.85%
39 Gemma 3 4B 95.04% 60.30% 68.57%
40 Claude Sonnet 4.6 94.66% 88.52% 91.15%
41 Claude 3.5 Sonnet 93.24% 76.75% 84.24%
42 Qwen 3.5 27B 92.55% 95.67% 90.85%
43 Ministral 3 14B 92.10% 79.03% 72.54%
44 DeepSeek V3.2 92.09% 81.58% 82.25%
45 Claude Sonnet 4.5 91.98% 83.78% 88.03%
46 Grok 4 Fast 89.92% 76.76% 86.15%
47 Aion 2.0 89.10% 90.91% 89.21%
48 DeepSeek V3 (2024-12-26) 88.56% 81.87% 83.68%
49 Grok 4 88.00% 89.67% 88.12%
50 GPT-4.1 86.01% 90.57% 88.68%
51 Llama 3.1 8B 85.64% 74.82% 63.37%
52 Gemini 2.5 Flash (Reasoning) 85.21% 82.25% 86.51%
53 Claude Sonnet 4 84.17% 84.02% 88.72%
54 DeepSeek V3.1 83.32% 76.65% 82.39%
55 ByteDance Seed 1.6 Flash 82.27% 84.16% 73.27%
56 Llama 3.1 70B 80.53% 81.03% 78.40%
57 Claude Opus 4 80.19% 88.81% 87.69%
58 Gemma 3 12B 80.06% 79.28% 78.41%
59 Hermes 3 405B 79.99% 69.02% 82.86%
60 Writer: Palmyra X5 79.93% 79.71% 79.57%
61 GPT-4o Mini (temp=1) 79.62% 82.16% 79.08%
62 GPT-4o Mini (temp=0) 79.49% 81.43% 78.29%
63 GPT-4.1 Mini 79.41% 82.30% 83.20%
64 Claude 3.5 Haiku 76.84% 82.57% 83.73%
65 Minimax M2.5 75.58% 90.42% 88.71%
66 GPT-4.1 Nano 75.17% 68.45% 71.94%
67 Ministral 3 3B 74.85% 72.38% 67.22%
68 Z.AI GLM 4.5 74.82% 79.19% 86.27%
69 GPT-4o, Aug. 6th (temp=1) 74.24% 82.44% 82.62%
70 Ministral 3 8B 74.12% 74.43% 71.76%
71 GPT-4o, May 13th (temp=1) 74.04% 80.69% 83.80%
72 GPT-4o, Aug. 6th (temp=0) 74.00% 82.11% 82.45%
73 Gemini 2.5 Flash Lite 73.32% 80.14% 81.08%
74 GPT-4o, May 13th (temp=0) 71.82% 83.13% 85.36%
75 WizardLM 2 8x22b 71.48% 67.14% 71.07%
76 Claude Haiku 4.5 70.28% 72.48% 85.14%
77 Qwen 2.5 72B 66.70% 76.43% 75.46%
78 Claude 3.7 Sonnet 66.35% 62.54% 83.39%
79 Gemini 2.5 Flash 64.79% 61.45% 80.60%
80 LFM2 24B 63.93% 69.48% 58.77%
81 Mistral Large 63.80% 73.04% 80.15%
82 Ministral 3B 62.17% 49.17% 61.29%
83 Claude 3 Haiku 61.88% 68.47% 71.19%
84 Cohere Command R+ (Aug. 2024) 60.43% 59.51% 69.03%
85 Ministral 8B 60.35% 46.82% 64.87%
86 Mistral Large 2 58.78% 69.19% 82.41%
87 Hermes 3 70B 58.14% 61.15% 72.57%
88 Arcee AI: Trinity Large (Preview) 57.16% 60.74% 73.33%
89 Rocinante 12B 53.12% 48.47% 54.55%
90 Mistral NeMO 47.57% 51.55% 65.04%
91 Arcee AI: Trinity Mini 43.97% 59.94% 70.90%