| Phi-2 | 45.2% | Self-reported | Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone via Epoch AI | — |
| Falcon-7B Instruct | 52.6% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Gemma 2B | 53.2% | Self-reported | Gemma: Open Models Based on Gemini Research and Technology via Epoch AI | — |
| MPT-7B (Instruct, Chat, StoryWriter) | 55.7% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Falcon-7B Instruct | 56.8% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| MPT-7B (Instruct, Chat, StoryWriter) | 59.6% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Falcon-7B Instruct | 61.1% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| MPT-7B (Instruct, Chat, StoryWriter) | 61.2% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| MPT-7B (Instruct, Chat, StoryWriter) | 61.6% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Mistral 7B v0.1 | 62.5% | Third party | Gemma: Open Models Based on Gemini Research and Technology via Epoch AI | — |
| Gemma 7B | 63.4% | Self-reported | Gemma: Open Models Based on Gemini Research and Technology via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 64.3% | Third party | Language Models are Few-Shot Learners via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 64.3% | Third party | GLaM: Efficient Scaling of Language Models with Mixture-of-Experts via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 64.3% | Third party | Language Models are Few-Shot Learners via Epoch AI | — |
| Falcon-7B Instruct | 64.6% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Llama 2 7B | 65.8% | Self-reported | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Llama 3 8B | 67.7% | Self-reported | Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone via Epoch AI | — |
| MPT-30B (Instruct, Chat) | 68.0% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 68.0% | Third party | Language Models are Few-Shot Learners via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 68.0% | Third party | GLaM: Efficient Scaling of Language Models with Mixture-of-Experts via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 68.0% | Third party | Language Models are Few-Shot Learners via Epoch AI | — |
| Llama 2 7B | 68.9% | Self-reported | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 71.2% | Third party | Language Models are Few-Shot Learners via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 71.2% | Third party | GLaM: Efficient Scaling of Language Models with Mixture-of-Experts via Epoch AI | — |
| text-davinci-001 (InstructGPT) | 71.2% | Third party | Language Models are Few-Shot Learners via Epoch AI | — |
| MPT-30B (Instruct, Chat) | 71.3% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Llama 2 7B | 72.1% | Third party | Gemma: Open Models Based on Gemini Research and Technology via Epoch AI | — |
| Llama 2 7B | 72.1% | Self-reported | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Gemma 7B | 72.3% | Self-reported | Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone via Epoch AI | — |
| Llama 2 13B | 73.1% | Self-reported | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| MPT-30B (Instruct, Chat) | 73.3% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| MPT-30B (Instruct, Chat) | 73.6% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Llama 2 7B | 73.7% | Self-reported | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Phi-3-medium | 73.9% | Self-reported | Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone via Epoch AI | — |
| Falcon-40B Instruct | 74.6% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Mistral 7B v0.1 | 75.2% | Self-reported | Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone via Epoch AI | — |
| Llama 2 13B | 77.2% | Self-reported | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Falcon-40B Instruct | 78.6% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Claude Instant 1.2 | 78.7% | Self-reported | Releasing Claude Instant 1.2 via Epoch AI | — |
| Claude Instant 1 | 78.9% | Self-reported | Model Card and Evaluations for Claude Models via Epoch AI | — |
| Claude Instant 1 | 78.9% | Self-reported | Releasing Claude Instant 1.2 via Epoch AI | — |
| Llama 2 13B | 79.4% | Self-reported | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Falcon-40B Instruct | 79.6% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Llama 2 13B | 79.6% | Third party | Gemma: Open Models Based on Gemini Research and Technology via Epoch AI | — |
| Llama 2 13B | 79.6% | Self-reported | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| Falcon-40B Instruct | 79.9% | Third party | Llama 2: Open Foundation and Fine-Tuned Chat Models via Epoch AI | — |
| DeepSeek-V2 | 80.0% | Self-reported | DeepSeek-V3 Technical Report via Epoch AI | — |
| Llama 3.1 405B | 82.7% | Self-reported | DeepSeek-V3 Technical Report via Epoch AI | — |
| DeepSeek-V3 | 82.9% | Self-reported | DeepSeek-V3 Technical Report via Epoch AI | — |
| GPT-4 | 84.8% | Third party | PapersWithCode via Epoch AI | — |
| GPT-3.5 Turbo | 85.8% | Self-reported | Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone via Epoch AI | — |
| Claude 1 | 86.7% | Self-reported | Model Card and Evaluations for Claude Models via Epoch AI | — |
| Claude 2 | 87.5% | Self-reported | Model Card and Evaluations for Claude Models via Epoch AI | — |