Ranking de LLM

Capacidades de grandes modelos de linguagem (dados de exemplo).

Atualizado:

Public ranking policy: rows are sorted by composite score (desc). Composite score is a weighted sum of normalized sub-metrics; ties are broken by higher recent activity.

Posição Modelo Provedor Tamanho Pontuação Notas
1 Xiaomi: MiMo-V2-Pro xiaomi 1.0M ctx 86.6 $2.00/1M avg · OpenRouter listing
2 Writer: Palmyra X5 writer 1.0M ctx 76.5 $3.30/1M avg · OpenRouter listing
3 Qwen: Qwen Plus 0728 (thinking) qwen 1.0M ctx 95.1 $0.52/1M avg · OpenRouter listing
4 MiniMax: MiniMax M1 minimax 1.0M ctx 89.3 $1.30/1M avg · OpenRouter listing
5 NVIDIA: Nemotron 3 Super (free) nvidia 262k ctx 62 $0.00/1M avg · OpenRouter listing
6 StepFun: Step 3.5 Flash stepfun 262k ctx 62 $0.20/1M avg · OpenRouter listing
7 Arcee AI: Trinity Large Thinking arcee-ai 262k ctx 62 $0.54/1M avg · OpenRouter listing
8 Mistral: Devstral 2 2512 mistralai 262k ctx 62 $1.20/1M avg · OpenRouter listing
9 MoonshotAI: Kimi K2 0905 moonshotai 262k ctx 62 $1.20/1M avg · OpenRouter listing
10 Morph: Morph V3 Large morph 262k ctx 62 $1.40/1M avg · OpenRouter listing
11 Kwaipilot: KAT-Coder-Pro V2 kwaipilot 256k ctx 62 $0.75/1M avg · OpenRouter listing
12 xAI: Grok Code Fast 1 x-ai 256k ctx 62 $0.85/1M avg · OpenRouter listing
13 Relace: Relace Apply 3 relace 256k ctx 62 $1.05/1M avg · OpenRouter listing
14 AI21: Jamba Large 1.7 ai21 256k ctx 62 $5.00/1M avg · OpenRouter listing
15 Cohere: Command A cohere 256k ctx 62 $6.25/1M avg · OpenRouter listing
16 Z.ai: GLM 4.6 z-ai 205k ctx 62 $1.15/1M avg · OpenRouter listing
17 DeepSeek: DeepSeek V3.2 deepseek 164k ctx 62 $0.32/1M avg · OpenRouter listing
18 TNG: DeepSeek R1T2 Chimera tngtech 164k ctx 62 $0.70/1M avg · OpenRouter listing
19 OpenAI: gpt-oss-120b (free) openai 131k ctx 62 $0.00/1M avg · OpenRouter listing
20 Meta: Llama 3.2 3B Instruct (free) meta-llama 131k ctx 62 $0.00/1M avg · OpenRouter listing