████████╗██╗   ██╗██████╗ ██╗███╗   ██╗ ██████╗ ███████╗ ██████╗ ██████╗ ██████╗ ███████╗    ██████╗ ██████╗ ███╗   ███╗
      ╚══██╔══╝██║   ██║██╔══██╗██║████╗  ██║██╔════╝ ██╔════╝██╔════╝██╔═══██╗██╔══██╗██╔════╝   ██╔════╝██╔═══██╗████╗ ████║
         ██║   ██║   ██║██████╔╝██║██╔██╗ ██║██║  ███╗███████╗██║     ██║   ██║██████╔╝█████╗     ██║     ██║   ██║██╔████╔██║
         ██║   ██║   ██║██╔══██╗██║██║╚██╗██║██║   ██║╚════██║██║     ██║   ██║██╔══██╗██╔══╝     ██║     ██║   ██║██║╚██╔╝██║
         ██║   ╚██████╔╝██║  ██║██║██║ ╚████║╚██████╔╝███████║╚██████╗╚██████╔╝██║  ██║███████╗██╗╚██████╗╚██████╔╝██║ ╚═╝ ██║
         ╚═╝    ╚═════╝ ╚═╝  ╚═╝╚═╝╚═╝  ╚═══╝ ╚═════╝ ╚══════╝ ╚═════╝ ╚═════╝ ╚═╝  ╚═╝╚══════╝╚═╝ ╚═════╝ ╚═════╝ ╚═╝     ╚═╝

Figure out which LLM is "best". Data from LMSYS's Chatbot Arena. Click here to learn more.

Large Language Models are AI program's that ingest massive amounts of text. By analyzing this data, they are able to figure out patterns and structures - they can translate texts; write poems, emails, newsletters, etc; answer questions, even strange and open question.
TuringScore.com keeps track of:
  • Model: The name of the model.
  • Elo: A relative "score" for each model. The higher the elo the better the model.
  • MMLU (Massive Multitask Language Understanding): A large benchmark that measures knowledge across 57 subjects like STEP, humanities, social sciences...
  • Licensing: Whether the model has weight's available and if they can be used commercially.

Model Elo MMLU Licensing
Gemini-1.5-Pro-Exp-0801 1299 Proprietary
GPT-4o-2024-05-13 1286 88.7 Proprietary
GPT-4o-mini-2024-07-18 1277 82 Proprietary
Claude 3.5 Sonnet 1271 88.7 Proprietary
Gemini Advanced App (2024-05-14) 1266 Proprietary
Meta-Llama-3.1-405b-Instruct 1264 88.6 Llama 3.1 Community
Gemini-1.5-Pro-001 1260 85.9 Proprietary
GPT-4-Turbo-2024-04-09 1257 Proprietary
Gemini-1.5-Pro-Preview-0409 1257 81.9 Proprietary
GPT-4-1106-preview 1251 Proprietary
Athene-70b 1249 CC-BY-NC-4.0
Mistral-Large-2407 1249 Mistral Research
Claude 3 Opus 1248 86.8 Proprietary
GPT-4-0125-preview 1245 Proprietary
Meta-Llama-3.1-70b-Instruct 1242 86 Llama 3.1 Community
Yi-Large-preview 1240 Proprietary
Gemini-1.5-Flash-001 1228 78.9 Proprietary
Reka-Core-20240722 1227 Proprietary
Deepseek-v2-API-0628 1221 DeepSeek
Gemma-2-27b-it 1218 Gemma license
Deepseek-Coder-v2-0724 1213 Proprietary
Yi-Large 1212 Proprietary
Gemini App (2024-01-24) 1209 Proprietary
Nemotron-4-340B-Instruct 1209 NVIDIA Open Model
GLM-4-0520 1207 Proprietary
Llama-3-70b-Instruct 1206 82 Llama 3 Community
Reka-Flash-20240722 1203 Proprietary
Claude 3 Sonnet 1201 79 Proprietary
Reka-Core-20240501 1199 83.2 Proprietary
Command R+ 1190 CC-BY-NC-4.0
Qwen2-72B-Instruct 1187 84.2 Qianwen LICENSE
Gemma-2-9b-it 1187 Gemma license
GPT-4-0314 1186 86.4 Proprietary
Qwen-Max-0428 1183 Proprietary
GLM-4-0116 1183 Proprietary
Claude 3 Haiku 1178 75.2 Proprietary
DeepSeek-Coder-V2-Instruct 1178 DeepSeek License
Reka-Flash-Preview-20240611 1165 Proprietary
GPT-4-0613 1162 Proprietary
Meta-Llama-3.1-8b-Instruct 1162 73 Llama 3.1 Community
Qwen1.5-110B-Chat 1161 80.4 Qianwen LICENSE
Mistral-Large-2402 1157 81.2 Proprietary
Yi-1.5-34B-Chat 1157 76.8 Apache-2.0
Reka-Flash-21B-online 1156 Proprietary
Llama-3-8b-Instruct 1152 68.4 Llama 3 Community
Claude-1 1149 77 Proprietary
Command R 1149 CC-BY-NC-4.0
Mistral Medium 1148 75.3 Proprietary
Qwen1.5-72B-Chat 1147 77.5 Qianwen LICENSE
Reka-Flash-21B 1147 73.5 Proprietary
Mixtral-8x22b-Instruct-v0.1 1146 77.8 Apache 2.0
Claude-2.0 1132 78.5 Proprietary
Gemini-1.0-Pro-001 1131 71.8 Proprietary
Zephyr-ORPO-141b-A35b-v0.1 1127 Apache 2.0
Gemma-2-2b-it 1127 51.3 Gemma license
Qwen1.5-32B-Chat 1125 73.4 Qianwen LICENSE
Mistral-Next 1124 Proprietary
Phi-3-Medium-4k-Instruct 1123 78 MIT
Starling-LM-7B-beta 1119 Apache-2.0
Claude-2.1 1118 Proprietary
GPT-3.5-Turbo-0613 1117 Proprietary
Mixtral-8x7b-Instruct-v0.1 1114 70.6 Apache 2.0
Claude-Instant-1 1111 73.4 Proprietary
Yi-34B-Chat 1111 73.5 Yi License
Gemini Pro 1111 71.8 Proprietary
Qwen1.5-14B-Chat 1109 67.6 Qianwen LICENSE
GPT-3.5-Turbo-0314 1106 70 Proprietary
WizardLM-70B-v1.0 1106 63.7 Llama 2 Community
GPT-3.5-Turbo-0125 1106 Proprietary
DBRX-Instruct-Preview 1103 73.7 DBRX LICENSE
Phi-3-Small-8k-Instruct 1101 75.7 MIT
Tulu-2-DPO-70B 1099 AI2 ImpACT Low-risk
Llama-2-70b-chat 1093 63 Llama 2 Community
Vicuna-33B 1091 59.2 Non-commercial
OpenChat-3.5-0106 1091 65.8 Apache-2.0
Snowflake Arctic Instruct 1090 67.3 Apache 2.0
Starling-LM-7B-alpha 1088 63.9 CC-BY-NC-4.0
Nous-Hermes-2-Mixtral-8x7B-DPO 1084 Apache-2.0
Gemma-1.1-7b-it 1083 64.3 Gemma license
NV-Llama2-70B-SteerLM-Chat 1081 68.5 Llama 2 Community
pplx-70b-online 1078 Proprietary
DeepSeek-LLM-67B-Chat 1077 71.3 DeepSeek License
OpenChat-3.5 1076 64.3 Apache-2.0
OpenHermes-2.5-Mistral-7b 1074 Apache-2.0
Mistral-7B-Instruct-v0.2 1072 Apache-2.0
Qwen1.5-7B-Chat 1070 61 Qianwen LICENSE
GPT-3.5-Turbo-1106 1068 Proprietary
Phi-3-Mini-4k-Instruct 1066 68.8 MIT
Phi-3-Mini-4k-Instruct-June-24 1064 70.9 MIT
Llama-2-13b-chat 1063 53.6 Llama 2 Community
SOLAR-10.7B-Instruct-v1.0 1062 66.2 CC-BY-NC-4.0
Dolphin-2.2.1-Mistral-7B 1062 Apache-2.0
WizardLM-13b-v1.2 1058 52.7 Llama 2 Community
Zephyr-7b-beta 1053 61.4 MIT
MPT-30B-chat 1045 50.4 CC-BY-NC-SA-4.0
pplx-7b-online 1045 Proprietary
Vicuna-13B 1042 55.8 Llama 2 Community
CodeLlama-34B-instruct 1042 53.7 Llama 2 Community
CodeLlama-70B-instruct 1042 Llama 2 Community
Zephyr-7b-alpha 1041 MIT
Llama-2-7b-chat 1037 45.8 Llama 2 Community
Gemma-7b-it 1037 64.3 Gemma license
Phi-3-Mini-128k-Instruct 1036 68.1 MIT
Qwen-14B-Chat 1035 66.5 Qianwen LICENSE
falcon-180b-chat 1034 68 Falcon-180B TII License
Guanaco-33B 1033 57.6 Non-commercial
Gemma-1.1-2b-it 1020 64.3 Gemma license
StripedHyena-Nous-7B 1017 Apache 2.0
OLMo-7B-instruct 1015 Apache-2.0
Mistral-7B-Instruct-v0.1 1008 55.4 Apache 2.0
Vicuna-7B 1005 49.8 Llama 2 Community
PaLM-Chat-Bison-001 1003 Proprietary
Qwen1.5-4B-Chat 989 56.1 Qianwen LICENSE
Gemma-2b-it 989 42.3 Gemma license
Koala-13B 964 44.7 Non-commercial
ChatGLM3-6B 956 Apache-2.0
GPT4All-13B-Snoozy 932 43 Non-commercial
MPT-7B-Chat 927 32 CC-BY-NC-SA-4.0
ChatGLM2-6B 924 45.5 Apache-2.0
RWKV-4-Raven-14B 922 25.6 Apache 2.0
Alpaca-13B 902 48.1 Non-commercial
OpenAssistant-Pythia-12B 893 27 Apache 2.0
ChatGLM-6B 879 36.1 Non-commercial
FastChat-T5-3B 868 47.7 Apache 2.0
StableLM-Tuned-Alpha-7B 840 24.4 CC-BY-NC-SA-4.0
Dolly-V2-12B 822 25.7 MIT
LLaMA-13B 798 47 Non-commercial
WizardLM-30B 58.7 Non-commercial
Vicuna-13B-16k 54.5 Llama 2 Community
WizardLM-13B-v1.1 50 Non-commercial
Tulu-30B 58.1 Non-commercial
Guanaco-65B 62.1 Non-commercial
OpenAssistant-LLaMA-30B 56 Non-commercial
WizardLM-13B-v1.0 52.3 Non-commercial
Vicuna-7B-16k 48.5 Llama 2 Community
Baize-v2-13B 48.9 Non-commercial
XGen-7B-8K-Inst 42.1 Non-commercial
Nous-Hermes-13B 49.3 Non-commercial
MPT-30B-Instruct 47.8 CC-BY-SA 3.0
Falcon-40B-Instruct 54.7 Apache 2.0
H2O-Oasst-OpenLLaMA-13B 42.8 Apache 2.0
LLaVA-v1.6-34B Apache 2.0
CogVLM2-llama3-chat-19b CogVLM2
InternVL2-26b MIT
GLM-4-AIR Proprietary
Snorkel-Mistral-PairRM-DPO Apache 2.0
The turing test (or imitation game) tries to distinguish machine from human.

© 2024 TuringScore

Feedback? Click here.