模型排行
137 个模型 · 1 个榜单 · 1 个数据来源
人气王座
开源之光
国产之光
最多人测
开源前三甲
llama-3.1-405b-instruct-fp8Meta · 总第 7 名1,252
llama-3.1-405b-instruct-bf16Meta · 总第 8 名1,251
mistral-large-2407Mistral · 总第 18 名1,230
| 排名 | 变化 | 模型 | 得分 | 置信区间 | 投票 | 价格 $/M |
|---|---|---|---|---|---|---|
| 1 | — | llama-3.1-405b-instruct-fp8开源Meta | 1,252 | ±3 | 52,089 | — |
| 2 | — | llama-3.1-405b-instruct-bf16开源Meta | 1,251 | ±4 | 14,596 | — |
| 3 | — | mistral-large-2407开源Mistral | 1,230 | ±3 | 48,270 | — |
| 4 | — | qwen-max-0919开源Alibaba | 1,230 | ±5 | 17,630 | — |
| 5 | 新 | athene-v2-chat开源NexusFlow | 1,229 | ±8 | 4,354 | — |
| 6 | ▼ 1 | deepseek-v2.5开源DeepSeek | 1,224 | ±4 | 26,023 | — |
| 7 | ▼ 1 | qwen2.5-72b-instruct开源Alibaba | 1,222 | ±4 | 26,302 | — |
| 8 | ▼ 1 | athene-70b-0725开源NexusFlow | 1,222 | ±5 | 20,689 | — |
| 9 | ▼ 1 | llama-3.1-nemotron-70b-instruct开源NVIDIA | 1,215 | ±7 | 7,270 | — |
| 10 | ▼ 1 | llama-3.1-70b-instruct开源Meta | 1,213 | ±3 | 47,975 | — |
| 11 | ▼ 1 | gemma-2-27b-it开源Google | 1,206 | ±3 | 57,595 | — |
| 12 | ▼ 1 | llama-3.1-nemotron-51b-instruct开源NVIDIA | 1,204 | ±9 | 3,942 | — |
| 13 | ▼ 1 | jamba-1.5-large开源AI21 Labs | 1,204 | ±6 | 9,178 | — |
| 14 | ▼ 1 | gemma-2-9b-it-simpo开源Princeton | 1,198 | ±5 | 10,609 | — |
| 15 | ▼ 1 | command-r-plus-08-2024开源Cohere | 1,197 | ±5 | 10,604 | — |
| 16 | ▼ 1 | nemotron-4-340b-instruct开源NVIDIA | 1,196 | ±3 | 20,637 | — |
| 17 | ▼ 1 | llama-3-70b-instruct开源Meta | 1,195 | ±2 | 163,859 | — |
| 18 | 新 | qwen2.5-coder-32b-instruct开源Alibaba | 1,191 | ±11 | 2,410 | — |
| 19 | ▼ 2 | command-r-plus开源Cohere | 1,183 | ±2 | 80,872 | — |
| 20 | 新 | c4ai-aya-expanse-32b开源Cohere | 1,182 | ±5 | 16,701 | — |
| 21 | ▼ 3 | deepseek-coder-v2开源DeepSeek | 1,181 | ±5 | 15,792 | — |
| 22 | ▼ 3 | qwen2-72b-instruct开源Alibaba | 1,181 | ±4 | 38,976 | — |
| 23 | ▼ 3 | gemma-2-9b-it开源Google | 1,177 | ±3 | 41,124 | — |
| 24 | ▼ 3 | command-r-08-2024开源Cohere | 1,175 | ±5 | 10,922 | — |
| 25 | ▼ 2 | ministral-8b-2410开源Mistral | 1,160 | ±9 | 4,587 | — |
| 26 | ▼ 4 | jamba-1.5-mini开源AI21 Labs | 1,159 | ±5 | 9,302 | — |
| 27 | ▼ 3 | qwen1.5-110b-chat开源Alibaba | 1,155 | ±4 | 27,472 | — |
| 28 | ▼ 3 | mixtral-8x22b-instruct-v0.1开源Mistral | 1,149 | ±3 | 53,819 | — |
| 29 | ▼ 3 | command-r开源Cohere | 1,148 | ±3 | 56,387 | — |
| 30 | ▼ 3 | qwen1.5-72b-chat开源Alibaba | 1,148 | ±3 | 40,618 | — |
| 31 | ▼ 3 | llama-3-8b-instruct开源Meta | 1,142 | ±2 | 109,293 | — |
| 32 | ▼ 3 | yi-1.5-34b-chat开源01 AI | 1,133 | ±4 | 25,142 | — |
| 33 | ▼ 2 | llama-3.1-8b-instruct开源Meta | 1,131 | ±3 | 46,435 | — |
| 34 | ▼ 4 | zephyr-orpo-141b-A35b-v0.1开源HuggingFace | 1,130 | ±9 | 4,854 | — |
| 35 | ▼ 3 | qwen1.5-32b-chat开源Alibaba | 1,126 | ±3 | 22,749 | — |
| 36 | ▼ 3 | phi-3-medium-4k-instruct开源Microsoft | 1,118 | ±4 | 26,140 | — |
| 37 | ▼ 3 | internlm2_5-20b-chat开源InternLM | 1,116 | ±5 | 10,710 | — |
| 38 | ▼ 3 | mixtral-8x7b-instruct-v0.1开源Mistral | 1,114 | ±0 | 76,133 | — |
| 39 | ▼ 3 | dbrx-instruct-preview开源Databricks | 1,113 | ±4 | 33,724 | — |
| 40 | ▼ 3 | gemma-2-2b-it开源Google | 1,113 | ±4 | 32,637 | — |
| 41 | 新 | granite-3.0-8b-instruct开源IBM | 1,110 | ±7 | 7,101 | — |
| 42 | ▼ 4 | qwen1.5-14b-chat开源Alibaba | 1,107 | ±4 | 18,663 | — |
| 43 | ▼ 4 | wizardlm-70b开源Microsoft | 1,104 | ±6 | 8,380 | — |
| 44 | ▼ 4 | snowflake-arctic-instruct开源Snowflake | 1,102 | ±3 | 34,150 | — |
| 45 | ▼ 4 | yi-34b-chat开源01 AI | 1,101 | ±5 | 15,925 | — |
| 46 | ▼ 4 | openchat-3.5-0106开源OpenChat · 另有 1 个档位 | 1,100 | ±4 | 12,968 | — |
| 47 | ▼ 4 | phi-3-small-8k-instruct开源Microsoft | 1,098 | ±4 | 18,504 | — |
| 48 | ▼ 4 | gemma-1.1-7b-it开源Google | 1,098 | ±4 | 25,071 | — |
| 49 | ▼ 4 | llama-3.2-3b-instruct开源Meta | 1,093 | ±7 | 8,448 | — |
| 50 | ▼ 4 | tulu-2-dpo-70b开源AllenAI/UW | 1,092 | ±6 | 6,662 | — |
智力进步
涨得最多 · Anthropic +57.2目前最高 · Anthropic 64.6最高与最低相差 18.1 分时间范围 2025年11月 – 2026年11月
64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi