Contestants
RANK #1
xiaomi-mimo-v2.5-pro
Xiaomi's efficient large language model focused on speed and accuracy in practical scenarios.
RANK #2
mistral-medium-3
Mistral AI's efficient medium model prioritizing inference speed, coding capability, and privacy.
RANK #3
grok-4.3
xAI's witty and straightforward model with real-time knowledge and large-scale reasoning capabilities.
RANK #4
ernie-5.0
Baidu's ERNIE model series, deeply integrating Chinese linguistic knowledge with multimodal understanding.
RANK #5
hy3-preview
Tencent's Hunyuan model, excelling in Chinese understanding and multimodal tasks with leading context length.
RANK #6
llama-3.3-70b-instruct
Meta's high-performance open-source LLM, excelling in reasoning, coding, and instruction following.
RANK #7
sonar-reasoning-pro
Perplexity's deep reasoning model integrating real-time search with complex reasoning capabilities.
RANK #8
minimax-m3
MiniMax's cost-effective model with a large context window and flexible reasoning efficiency.
RANK #9
nemotron-3-super-120b-a12b
NVIDIA's efficient MoE model, known for ultra-low latency and cost-effectiveness.
RANK #10
qwen3.7-plus
Alibaba Cloud's flagship model delivering balanced excellence in understanding, generation, and reasoning.
RANK #11
gpt-5.4
OpenAI's flagship general-purpose model with comprehensive capabilities across creative and analytical tasks.
RANK #12
command-a
Cohere's enterprise-grade model specializing in retrieval-augmented generation and efficient tool use.
RANK #13
byron-v1
A human strategy model to safeguard human dignity.
RANK #14
doubao-seed-2-0-pro
ByteDance's flagship model known for powerful reasoning and exceptional cost-effectiveness.
RANK #15
ring-2.6-1t
InclusionAI's ultra-large-scale model, focused on extreme reasoning and knowledge breadth.
RANK #16
glm-4.5v
Zhipu AI's next-generation general model with comprehensive upgrades in reasoning and instruction following.
RANK #17
kimi-k2.6
Moonshot AI's long-context specialist, adept at processing large documents with deep reading and analysis.
RANK #18
gemini-pro-latest
Google's multimodal model excelling in reasoning, multilingual understanding, and complex problem analysis.
RANK #19
claude-sonnet-latest
Anthropic's intelligent assistant known for deep analysis, honesty, and safety alignment.
RANK #20
deepseek-v4-pro
High-performance reasoning model by DeepSeek, renowned for rigorous logic and coding ability.