Language
9 scenarios across 2 subcategories. 118 models scored.
Subcategories
| Subcategory | Avg Score | Best Model | Best Score |
|---|---|---|---|
| Comprehension | 81.86% | Claude Opus 4.6 (Reasoning) | 100.00% |
| Generation | 85.28% | Claude Sonnet 4.6 | 100.00% |
Model Leaderboard
All models ranked by their Language category score.
| # | Model | Language | Comprehension | Generation | Overall |
|---|---|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 100.00% | 100.00% | 100.00% | 91.15% |
| 2 | DeepSeek-V2 Chat | 100.00% | 100.00% | 100.00% | 84.83% |
| 3 | Z.AI GLM 5 Turbo | 99.90% | 100.00% | 99.80% | 94.27% |
| 4 | Claude Opus 4.5 | 99.66% | 100.00% | 99.31% | 89.69% |
| 5 | Hermes 3 405B | 99.57% | 100.00% | 99.14% | 82.86% |
| 6 | Grok 4.20 (Beta, Reasoning) | 99.08% | 100.00% | 98.15% | 91.49% |
| 7 | GPT-4o, May 13th (temp=0) | 98.72% | 100.00% | 97.44% | 85.36% |
| 8 | GPT-5.4 Mini (Reasoning) | 98.12% | 100.00% | 96.25% | 90.65% |
| 9 | Claude Sonnet 4.6 (Reasoning) | 97.58% | 100.00% | 95.15% | 93.66% |
| 10 | Z.AI GLM 4.5 | 97.33% | 95.00% | 99.67% | 86.27% |
| 11 | MoonshotAI: Kimi K2.5 | 97.10% | 100.00% | 94.19% | 91.04% |
| 12 | DeepSeek V3.1 | 96.87% | 95.00% | 98.74% | 82.39% |
| 13 | ByteDance Seed 2.0 Lite | 96.80% | 100.00% | 93.60% | 84.80% |
| 14 | Z.AI GLM 4.6 | 96.60% | 100.00% | 93.20% | 89.11% |
| 15 | GPT-5 Mini | 96.49% | 95.00% | 97.98% | 92.62% |
| 16 | Aion 2.0 | 96.17% | 100.00% | 92.34% | 89.21% |
| 17 | Claude Opus 4.6 | 96.13% | 100.00% | 92.27% | 92.35% |
| 18 | Claude Opus 4.6 (Reasoning) | 96.12% | 100.00% | 92.23% | 95.02% |
| 19 | MiniMax M2.5 | 96.05% | 95.00% | 97.10% | 88.71% |
| 20 | ByteDance Seed 1.6 | 95.63% | 100.00% | 91.26% | 90.70% |
| 21 | Qwen 3.5 27B | 95.52% | 100.00% | 91.05% | 90.85% |
| 22 | Qwen 3.5 Plus (2026-02-15) | 95.10% | 100.00% | 90.21% | 85.96% |
| 23 | Qwen 3.5 122B | 95.01% | 100.00% | 90.03% | 91.53% |
| 24 | Qwen 3.5 397B A17B | 95.01% | 100.00% | 90.03% | 91.73% |
| 25 | Gemini 3 Flash (Preview) | 95.00% | 90.00% | 100.00% | 85.35% |
| 26 | Gemini 3.1 Flash Lite (Preview) | 94.98% | 95.00% | 94.95% | 85.87% |
| 27 | Gemini 3 Flash (Preview, Reasoning) | 94.93% | 95.00% | 94.87% | 90.50% |
| 28 | GPT-5.4 (Reasoning) | 94.90% | 95.00% | 94.80% | 93.24% |
| 29 | Gemini 3.1 Pro (Preview) | 94.90% | 95.00% | 94.80% | 94.37% |
| 30 | GPT-4.1 | 93.91% | 90.00% | 97.82% | 88.68% |
| 31 | GPT-5.1 | 93.64% | 90.00% | 97.28% | 92.54% |
| 32 | Stealth: Hunter Alpha | 93.35% | 95.00% | 91.70% | 87.34% |
| 33 | Claude Opus 4 | 93.01% | 100.00% | 86.02% | 87.69% |
| 34 | Claude 3.7 Sonnet | 92.95% | 90.00% | 95.89% | 83.39% |
| 35 | Gemini 2.5 Pro | 92.57% | 90.00% | 95.14% | 88.53% |
| 36 | GPT-4o, May 13th (temp=1) | 92.52% | 95.00% | 90.04% | 83.80% |
| 37 | Stealth: Aurora Alpha | 92.50% | 85.00% | 100.00% | 83.79% |
| 38 | GPT-5.4 Mini (Reasoning, Low) | 92.45% | 85.00% | 99.89% | 85.75% |
| 39 | Claude Sonnet 4.5 | 92.39% | 95.00% | 89.77% | 88.03% |
| 40 | Z.AI GLM 5 | 92.06% | 95.00% | 89.12% | 91.23% |
| 41 | Mistral Large 3 | 92.02% | 100.00% | 84.05% | 85.43% |
| 42 | Qwen 3.5 35B | 91.95% | 100.00% | 83.90% | 88.00% |
| 43 | Qwen 3.5 Flash | 91.94% | 95.00% | 88.89% | 86.38% |
| 44 | Claude Haiku 4.5 | 91.84% | 95.00% | 88.67% | 85.14% |
| 45 | GPT-5 | 91.50% | 85.00% | 98.01% | 91.93% |
| 46 | Claude Sonnet 4 | 91.31% | 90.00% | 92.61% | 88.72% |
| 47 | GPT-5.2 | 91.19% | 85.00% | 97.39% | 90.26% |
| 48 | Grok 4.20 (Beta) | 91.17% | 85.00% | 97.34% | 83.85% |
| 49 | GPT-5.4 (Reasoning, Low) | 90.79% | 85.00% | 96.58% | 91.41% |
| 50 | Grok 4 | 90.61% | 85.00% | 96.22% | 88.12% |
| 51 | ByteDance Seed 2.0 Mini | 90.12% | 90.00% | 90.24% | 86.91% |
| 52 | Gemini 3 Pro (Preview) | 89.64% | 95.00% | 84.29% | 88.79% |
| 53 | GPT-4.1 Mini | 89.64% | 80.00% | 99.27% | 83.20% |
| 54 | Grok 4.1 Fast | 88.76% | 90.00% | 87.51% | 89.55% |
| 55 | GPT-5.4 Mini | 88.75% | 80.00% | 97.50% | 82.43% |
| 56 | Mistral Large | 88.64% | 90.00% | 87.27% | 80.15% |
| 57 | Stealth: Healer Alpha | 88.45% | 90.00% | 86.90% | 85.93% |
| 58 | Qwen 3.5 9B | 88.18% | 90.00% | 86.36% | 86.05% |
| 59 | DeepSeek V3 (2024-12-26) | 87.88% | 100.00% | 75.76% | 83.68% |
| 60 | Z.AI GLM 4.7 Flash | 87.67% | 95.00% | 80.35% | 84.82% |
| 61 | Nemotron 3 Nano | 87.63% | 80.00% | 95.26% | 77.73% |
| 62 | Inception Mercury 2 | 87.32% | 75.00% | 99.64% | 83.85% |
| 63 | DeepSeek V3 (2025-03-24) | 86.42% | 100.00% | 72.83% | 81.99% |
| 64 | Gemini 2.5 Flash | 86.23% | 75.00% | 97.46% | 80.60% |
| 65 | Gemini 2.5 Flash (Reasoning) | 86.06% | 75.00% | 97.13% | 86.51% |
| 66 | Claude 3.5 Sonnet | 85.62% | 75.00% | 96.24% | 84.24% |
| 67 | Z.AI GLM 4.7 | 85.46% | 80.00% | 90.93% | 88.69% |
| 68 | Mistral Large 2 | 85.22% | 100.00% | 70.43% | 82.41% |
| 69 | DeepSeek V3.2 | 85.01% | 80.00% | 90.03% | 82.25% |
| 70 | MiniMax M2.7 | 84.80% | 100.00% | 69.59% | 89.10% |
| 71 | Grok 4 Fast | 84.61% | 85.00% | 84.22% | 86.15% |
| 72 | Qwen 3 32B | 84.61% | 90.00% | 79.21% | 82.21% |
| 73 | GPT-5.4 Nano (Reasoning) | 83.99% | 70.00% | 97.99% | 81.36% |
| 74 | Gemini 2.5 Flash Lite | 82.75% | 80.00% | 85.50% | 81.08% |
| 75 | GPT-4o, Aug. 6th (temp=1) | 82.21% | 65.00% | 99.43% | 82.62% |
| 76 | Claude 3.5 Haiku | 82.12% | 90.00% | 74.25% | 83.73% |
| 77 | GPT-5.4 Nano (Reasoning, Low) | 81.87% | 70.00% | 93.74% | 79.48% |
| 78 | Hermes 3 70B | 81.66% | 75.00% | 88.33% | 72.57% |
| 79 | GPT-5.4 | 81.49% | 65.00% | 97.98% | 84.32% |
| 80 | Nemotron 3 Super | 81.41% | 65.00% | 97.82% | 84.56% |
| 81 | GPT-5.4 Nano | 80.82% | 70.00% | 91.64% | 74.40% |
| 82 | Mistral NeMO | 80.80% | 95.00% | 66.60% | 65.04% |
| 83 | Inception Mercury | 80.37% | 65.00% | 95.74% | 79.50% |
| 84 | Llama 3.1 70B | 80.18% | 75.00% | 85.36% | 78.40% |
| 85 | Gemma 3 12B | 80.10% | 70.00% | 90.19% | 78.41% |
| 86 | o4 Mini | 80.00% | 60.00% | 100.00% | 88.35% |
| 87 | o4 Mini High | 79.76% | 60.00% | 99.51% | 90.29% |
| 88 | GPT-4.1 Nano | 78.95% | 65.00% | 92.91% | 71.94% |
| 89 | Arcee AI: Trinity Large (Preview) | 78.38% | 80.00% | 76.76% | 73.33% |
| 90 | WizardLM 2 8x22b | 78.05% | 95.00% | 61.10% | 71.07% |
| 91 | GPT-4o Mini (temp=1) | 77.50% | 55.00% | 100.00% | 79.08% |
| 92 | Gemma 3 27B | 77.21% | 65.00% | 89.42% | 77.85% |
| 93 | GPT-5 Nano | 77.18% | 55.00% | 99.35% | 82.60% |
| 94 | GPT-4o, Aug. 6th (temp=0) | 75.00% | 50.00% | 100.00% | 82.45% |
| 95 | GPT-4o Mini (temp=0) | 75.00% | 50.00% | 100.00% | 78.29% |
| 96 | Gemini 2.5 Flash Lite (Reasoning) | 74.36% | 65.00% | 83.71% | 85.75% |
| 97 | Mistral Small 3.2 24B | 72.77% | 75.00% | 70.53% | 78.60% |
| 98 | Claude 3 Haiku | 72.76% | 65.00% | 80.53% | 71.19% |
| 99 | Gemma 3 4B | 72.28% | 70.00% | 74.56% | 68.57% |
| 100 | Arcee AI: Trinity Mini | 70.59% | 60.00% | 81.18% | 70.90% |
| 101 | Qwen 2.5 72B | 68.95% | 70.00% | 67.91% | 75.46% |
| 102 | Ministral 3 3B | 68.10% | 100.00% | 36.20% | 67.22% |
| 103 | Cohere Command R+ (Aug. 2024) | 66.58% | 60.00% | 73.15% | 69.03% |
| 104 | LFM2 24B | 64.64% | 65.00% | 64.28% | 58.77% |
| 105 | Llama 3.1 8B | 64.06% | 55.00% | 73.13% | 63.37% |
| 106 | Rocinante 12B | 63.45% | 75.00% | 51.90% | 54.55% |
| 107 | ByteDance Seed 1.6 Flash | 61.23% | 40.00% | 82.45% | 73.27% |
| 108 | Qwen3 235B A22B Instruct 2507 | 60.83% | 75.00% | 46.67% | 80.10% |
| 109 | Mistral Small 4 (Reasoning) | 60.53% | 50.00% | 71.06% | 82.39% |
| 110 | Writer: Palmyra X5 | 56.58% | 70.00% | 43.16% | 79.57% |
| 111 | Ministral 8B | 53.91% | 55.00% | 52.81% | 64.87% |
| 112 | Mistral Small 4 | 51.96% | 55.00% | 48.93% | 76.46% |
| 113 | Mistral Medium 3.1 | 49.50% | 50.00% | 49.00% | 77.83% |
| 114 | Ministral 3 8B | 48.96% | 50.00% | 47.92% | 71.76% |
| 115 | Llama 3.1 Nemotron 70B | 46.80% | 60.00% | 33.61% | 74.70% |
| 116 | Ministral 3B | 42.25% | 25.00% | 59.49% | 61.29% |
| 117 | Mistral Small Creative | 41.85% | 50.00% | 33.69% | 73.27% |
| 118 | Ministral 3 14B | 30.00% | 50.00% | 10.00% | 72.54% |