Global LLM Tracking

Real-time updates on global mainstream LLM dynamics, providing authoritative evaluation rankings and parameter comparisons

Model Capability Leaderboard (Top 5)

Rank Model Name Developer Context Window MMLU Score Release Date
#1 Claude 3 Opus Anthropic 200k 86.8 2024-03
#2 GPT-4 Turbo OpenAI 128k 86.4 2023-11
#3 Gemini 1.5 Pro Google 1M+ 85.9 2024-02
#4 Llama 3 70B Meta 8k 82.0 2024-04
#5 Qwen1.5-72B Alibaba 32k 77.5 2024-02

Model Categories

Latest Releases

๐Ÿฆ™
Open Weights

Llama 3

Meta's latest open-source large model, with 8B and 70B versions, outperforming peer models in multiple benchmark tests.

8B/70B Params Meta 2024-04
๐Ÿง 
Proprietary

GPT-4o

OpenAI's latest all-purpose model with native multimodal capabilities, real-time voice interaction, faster speed, and lower price.

Multimodal OpenAI 2024-05
๐Ÿ‡ซ๐Ÿ‡ท
Open Weights

Mixtral 8x22B

Mistral AI's latest MoE model with 141B parameters, activating 39B, with strong performance.

MoE Mistral 2024-04
๐Ÿง 
Proprietary

Claude 3 Opus

Anthropic's strongest model, surpassing GPT-4 in multiple evaluations, with a 200k context window.

200K Context Anthropic 2024-03
๐Ÿ”ฎ
Proprietary

Gemini 1.5 Pro

Google's latest model supporting million-level context windows, excelling in long document understanding.

1M Context Google 2024-02
๐Ÿฎ
Open Weights

Qwen1.5-72B

Alibaba's Tongyi Qianwen latest version, excelling in Chinese understanding and generation, with strong multilingual capabilities.

32K Context Alibaba 2024-02