Ranking de LLMs — Comparação de modelos de IA da OpenAI, Anthropic, Google, SpaceXAI e outras
Comparação e ranking do desempenho de mais de 250 modelos de IA (LLMs) em métricas importantes como inteligência, preço, desempenho e velocidade (velocidade de saída — tokens por segundo — e latência — TTFT), janela de contexto e outras.
Para mais detalhes, inclusive sobre nossa metodologia, consulte as perguntas frequentes.
Inteligência
UpdatedClaude Fable 5.1 (max with fallback) e Claude Fable 5.1 (xhigh with fallback) são os modelos com maior inteligência, seguidos por GPT-6 Astra (max) e GPT-6 Astra (xhigh).
Velocidade de saída
Celeris-1 e Mercury 2 são os modelos mais rápidos, seguidos por Gemini 2.5 Flash-Lite e Gemini 3.5 Flash-Lite.
Latência
Gemini 2.5 Flash-Lite (Non-reasoning) e Command A+ são os modelos com menor latência, seguidos por North Mini Code e Gemini 2.5 Flash (Non-reasoning).
Custo por tarefa
Granite 4.2 3B e GPT-5.6 Luna (low) têm o menor custo por tarefa, seguidos por gpt-oss-20b (high) e GPT-5.6 Luna (medium).
Janela de contexto
Llama 4 Scout e Grok 4.20 0309 permitem as maiores janelas de contexto, seguidos por Gemini 1.5 Pro (May) e Grok 4.1 Fast.
Análises adicionais | ||||||||
|---|---|---|---|---|---|---|---|---|
Claude Fable 5.1 (max with fallback) | 1M | Anthropic | 53 | $7.63 | 65 | 213.23 | 220.91 | |
Claude Fable 5.1 (xhigh with fallback) | 1M | Anthropic | 53 | $5.98 | 60 | 96.81 | 105.19 | |
GPT-6 Astra (max) | 1M | OpenAI | 53 | $3.26 | 60 | 334.04 | 342.40 | |
GPT-6 Astra (xhigh) | 1M | OpenAI | 53 | $2.31 | 55 | 123.68 | 132.75 | |
Claude Fable 5.1 (high with fallback) | 1M | Anthropic | 51 | $3.91 | 54 | 13.19 | 22.38 | |
GPT-6 Astra (high) | 1M | OpenAI | 51 | $1.72 | 55 | 36.95 | 46.07 | |
Claude Opus 5 (max) | 1M | Anthropic | 51 | $5.86 | 52 | 46.50 | 56.18 | |
Claude Fable 5 (with fallback) | 1M | Anthropic | 50 | $8.75 | 65 | 101.16 | 108.91 | |
GPT-6 Astra (medium) | 1M | OpenAI | 50 | $1.54 | 55 | 4.01 | 13.08 | |
Claude Opus 5 (xhigh) | 1M | Anthropic | 50 | $4.88 | 49 | 18.47 | 28.58 | |
Claude Fable 5.1 (medium with fallback) | 1M | Anthropic | 49 | $2.98 | 51 | 8.03 | 17.87 | |
Claude Opus 5 (high) | 1M | Anthropic | 48 | $3.61 | 50 | 10.43 | 20.34 | |
Muse Spark 1.3 (max) | 1M | Meta | 48 | $1.60 | 228 | 22.23 | 33.22 | |
GPT-5.6 Sol (max) | 1M | OpenAI | 47 | $1.99 | 64 | 110.94 | 118.72 | |
Claude Fable 5.1 (low with fallback) | 1M | Anthropic | 47 | $2.37 | 50 | 4.87 | 14.78 | |
GPT-6 Astra (low) | 1M | OpenAI | 46 | $0.82 | 51 | 2.66 | 12.40 | |
Muse Spark 1.3 (xhigh) | 1M | Meta | 45 | $1.37 | 234 | 22.12 | 32.79 | |
Claude Opus 5 (medium) | 1M | Anthropic | 45 | $2.19 | 50 | 3.59 | 13.60 | |
GLM-5.3 (max) | 1M |  Z AI | 45 | $2.01 | 73 | 3.08 | 37.34 | |
Grok 4.6 (high) | 500k | SpaceXAI | 44 | $1.86 | 60 | 40.56 | 48.84 | |
Grok 4.6 (xhigh) | 500k | SpaceXAI | 44 | $2.32 | 56 | 41.74 | 50.68 | |
GPT-5.6 Sol (xhigh) | 1M | OpenAI | 44 | $1.18 | 58 | 38.98 | 47.64 | |
Kimi K3 (max) | 1.05M |  Kimi | 44 | $2.00 | 35 | 4.46 | 75.74 | |
Grok 4.6 (medium) | 500k | SpaceXAI | 43 | $1.50 | 54 | 34.30 | 43.52 | |
GPT-5.6 Sol (high) | 1M | OpenAI | 42 | $0.81 | 57 | 9.69 | 18.51 | |
GPT-5.6 Terra (max) | 1M | OpenAI | 42 | $1.40 | 113 | 153.26 | 157.70 | |
GLM-5.3-Flash | 1M | Z AI | 42 | $0.25 | 114 | 2.45 | 24.33 | |
Gemini 3.8 Flash (high) | 1M | Google | 41 | $1.24 | 303 | 16.28 | 17.93 | |
Qwen3.8 Max | 1M |  Alibaba | 40 | $2.67 | 41 | 2.64 | 64.07 | |
Qwen3.8 2.4T A95B | 984k | Alibaba | 40 | $2.16 | 41 | 2.80 | 63.49 | |
Gemini 3.8 Flash (medium) | 1M | Google | 40 | $0.93 | -- | -- | -- | |
Qwen3.8-Flash-Next | 256k | Alibaba | 40 | $0.37 | 53 | 2.73 | 50.07 | |
Muse Spark 1.2 (xhigh) | 1.05M | Meta | 40 | $0.97 | 181 | 18.23 | 32.03 | |
Claude Opus 5 (low) | 1M | Anthropic | 40 | $1.10 | 47 | 2.73 | 13.46 | |
Gemini 3.7 Flash (medium) | 1M | Google | 40* | -- | 279 | 4.69 | 6.49 | |
DeepSeek V4.1 Flash (max) | 1M |  DeepSeek | 40 | $0.27 | 214 | 1.12 | 12.78 | |
GPT-5.6 Sol (medium) | 1M | OpenAI | 39 | $0.50 | 58 | 3.11 | 11.79 | |
Gemini 3.7 Flash (high) | 1M | Google | 39 | $0.93 | 303 | 9.81 | 11.46 | |
Grok 4.5 (high) | 500k | SpaceXAI | 39 | $1.04 | 60 | 6.84 | 15.24 | |
Claude Sonnet 5 (max) | 1M | Anthropic | 38 | $5.09 | 76 | 195.83 | 202.39 | |
GPT-5.6 Terra (xhigh) | 1M | OpenAI | 38 | $0.63 | 101 | 21.09 | 26.06 | |
GPT-5.6 Luna (max) | 1M | OpenAI | 38 | $0.18 | 115 | 128.28 | 132.62 | |
Gemini 3.7 Flash (low) | 1M | Google | 37* | -- | 313 | 0.81 | 2.41 | |
DeepSeek V4 Pro 0813 (max) | 1M | DeepSeek | 36 | $0.67 | 91 | 1.69 | 29.10 | |
Agnes 3.0 Flash | 1M |  Sapiens AI | 36* | -- | 237 | 1.84 | 12.40 | |
Grok 4.6 (low) | 500k | SpaceXAI | 35 | $0.48 | 53 | 5.90 | 15.26 | |
Agnes 2.5 Pro Beta | 1M | Sapiens AI | 35* | -- | -- | -- | -- | |
DeepSeek V4 Flash Vision (max) | 1M | DeepSeek | 35 | $0.31 | 214 | 1.32 | 12.99 | |
GPT-5.6 Luna (xhigh) | 1M | OpenAI | 35 | $0.09 | 118 | 49.03 | 53.28 | |
Claude Sonnet 5 (xhigh) | 1M | Anthropic | 35 | $2.87 | 66 | 23.62 | 31.22 | |
DeepSeek V4 Flash 0731 (max) | 1M | DeepSeek | 35 | $0.22 | 214 | 1.43 | 13.11 | |
GPT-5.6 Terra (high) | 1M | OpenAI | 34 | $0.34 | 92 | 2.26 | 7.68 | |
Gemini 3.6 Flash | 1M | Google | 34 | $0.93 | 183 | 18.73 | 21.45 | |
GLM-5.2 (max) | 1M | Z AI | 34 | $0.96 | 75 | 8.00 | 41.43 | |
Qwen3.8 27B (xhigh) | 256k | Alibaba | 34 | $0.82 | 45 | 3.88 | 59.65 | |
GPT-5.6 Sol (low) | 1M | OpenAI | 34 | $0.26 | 58 | 2.87 | 11.56 | |
Gemini 3.8 Flash (low) | 1M | Google | 34 | -- | -- | -- | -- | |
Gemini 3.5 Flash (medium) | 1M | Google | 34* | -- | 223 | 13.39 | 15.64 | |
Motif 3 | 262k |  Motif Technologies | 34* | -- | -- | -- | -- | |
GPT-5.3 Codex (xhigh) | 400k | OpenAI | 33* | -- | 123 | 40.73 | 44.81 | |
GPT-5.6 Luna (high) | 1M | OpenAI | 32 | $0.04 | 109 | 8.57 | 13.13 | |
Motif 3 (Beta) | 262k | Motif Technologies | 32* | -- | -- | -- | -- | |
Claude Sonnet 5 (high) | 1M | Anthropic | 32 | $1.79 | 60 | 2.76 | 11.06 | |
DeepSeek V4 Pro (max) | 1M | DeepSeek | 31 | $0.12 | 86 | 1.63 | 58.00 | |
K2 Horizon 375B A23B | 524k |  MBZUAI Institute of Foundation Models | 31 | -- | -- | -- | -- | |
Kimi K3 (low) | 1.05M | Kimi | 30 | $1.15 | 35 | 3.94 | 74.74 | |
GPT-5.6 Terra (medium) | 1M | OpenAI | 30 | $0.18 | 96 | 1.72 | 6.91 | |
Apodex 1.1 | 256k |  Apodex | 30* | -- | -- | -- | -- | |
Gemini 3.1 Pro Preview | 1M | Google | 30 | $0.67 | 110 | 30.58 | 35.13 | |
DeepSeek V4 Pro (high) | 1M | DeepSeek | 30* | -- | 92 | 1.75 | 28.85 | |
MiniMax-M3 | 1M |  MiniMax | 30 | $0.51 | 88 | 1.35 | 29.85 | |
Claude Sonnet 5 (Non-reasoning) | 1M | Anthropic | 29* | -- | 61 | 0.72 | 8.87 | |
Claude Sonnet 5 (medium) | 1M | Anthropic | 28 | $1.00 | 57 | 1.18 | 9.90 | |
GPT-5.6 Sol (Non-reasoning) | 1M | OpenAI | 28* | -- | 62 | 1.19 | 9.26 | |
Nex-N2-Pro | 262k |  Nex AGI | 28* | -- | -- | -- | -- | |
Solar Pro 4 | 512k |  Upstage | 28* | -- | 75 | 1.93 | 35.33 | |
GPT-5.6 Terra (low) | 1M | OpenAI | 28 | $0.14 | 95 | 1.58 | 6.86 | |
Qwen3.8 27B (medium) | 256k | Alibaba | 28 | $0.90 | 51 | 3.91 | 52.67 | |
Agnes 2.5 Pro Alpha | 1M | Sapiens AI | 28 | -- | 172 | 4.02 | 18.55 | |
JT-4.1 Flash 236B A21B | 256k |  China Mobile | 27* | -- | -- | -- | -- | |
Quasar 438B (max) | 1M |  Multiverse Computing | 27 | $2.02 | 171 | 1.02 | 15.62 | |
GPT-5.5 Instant (June 2026) | 400k | OpenAI | 27 | $0.69 | 131 | 1.21 | 20.25 | |
Qwen3.8 27B (low) | 256k | Alibaba | 26 | $0.83 | 50 | 4.02 | 54.31 | |
MiMo-V2.5-Pro | 1M |  Xiaomi | 26 | $0.05 | 39 | 5.92 | 69.92 | |
Kimi K2.7 Code | 256k | Kimi | 26 | $0.54 | 48 | 2.89 | 59.21 | |
Inkling Small | 1M |  Thinking Machines | 26 | $0.09 | 178 | 2.13 | 16.15 | |
Qwen3.7 Plus | 1M | Alibaba | 26 | $0.33 | 70 | 2.22 | 37.77 | |
Hy3 | 256k |  Tencent | 26 | $0.07 | 91 | 2.69 | 30.25 | |
K2 Horizon MoVA 36B A4B | 524k | MBZUAI Institute of Foundation Models | 26 | -- | -- | -- | -- | |
Inkling | 1M | Thinking Machines | 26 | $0.61 | 90 | 2.86 | 30.55 | |
GPT-5.6 Luna (medium) | 1M | OpenAI | 25 | $0.02 | 107 | 2.39 | 7.04 | |
MiMo-V2-Omni-0327 | 256k | Xiaomi | 25* | -- | -- | -- | -- | |
Ling-3.0-flash-VL | 262k |  InclusionAI | 25 | -- | 144 | 2.26 | 19.60 | |
Grok 4.3 (medium) | 1M | SpaceXAI | 25* | -- | 109 | 11.15 | 15.74 | |
Claude Sonnet 5 (low) | 1M | Anthropic | 25 | $0.51 | 56 | 1.18 | 10.14 | |
Solar Open2 250B | 1.05M | Upstage | 25* | -- | -- | -- | -- | |
Grok 4.3 (low) | 1M | SpaceXAI | 24* | -- | 112 | 4.87 | 9.34 | |
MiMo-V2-Omni | 256k | Xiaomi | 24* | -- | -- | -- | -- | |
Gemini 3.5 Flash (minimal) | 1M | Google | 24* | -- | 205 | 0.82 | 3.26 | |
Nemotron 3 Ultra | 262k |  NVIDIA | 23 | $0.58 | 204 | 1.83 | 15.44 | |
Claude Sonnet 4.6 (Non-reasoning, low) | 1M | Anthropic | 23* | -- | 41 | 1.33 | 13.58 | |
A.X-K2 | 262k |  SK Telecom | 23* | -- | -- | -- | -- | |
Gemini 3.5 Flash-Lite | 1M | Google | 23 | $0.12 | 366 | 7.86 | 9.23 | |
MiMo-V2-Flash (Feb 2026) | 256k | Xiaomi | 22* | -- | -- | -- | -- | |
GLM-5.2 (Non-reasoning) | 1M | Z AI | 22* | -- | 160 | 1.57 | 4.70 | |
Qwen3.8 27B | 256k | Alibaba | 22* | -- | 51 | 4.01 | 13.81 | |
MiMo-V2.5 | 1M | Xiaomi | 22 | $0.02 | 45 | 7.05 | 62.76 | |
GPT-5.6 Terra (Non-reasoning) | 1M | OpenAI | 22* | -- | 93 | 0.86 | 6.26 | |
Qwen3.6 27B | 262k | Alibaba | 22 | $0.62 | 57 | 3.66 | 111.22 | |
G9v3-39A5B | 131k |  AI9Stars | 22* | -- | -- | -- | -- | |
KAT-Coder-Pro V2 | 256k |  KwaiKAT | 22* | -- | -- | -- | -- | |
GPT-5.6 Luna (low) | 1M | OpenAI | 22 | $0.01 | 101 | 2.14 | 7.07 | |
Qwen3.5 397B A17B (Non-reasoning) | 262k | Alibaba | 21* | -- | 86 | 2.16 | 7.95 | |
K2 Horizon 7B | 524k | MBZUAI Institute of Foundation Models | 21 | -- | -- | -- | -- | |
DeepSeek V4 Pro (Non-reasoning) | 1M | DeepSeek | 21* | -- | 87 | 1.57 | 7.29 | |
Ling 3.0 Flash | 262k | InclusionAI | 21 | $0.03 | 332 | 2.73 | 10.26 | |
Qwen3.5 Omni Plus | 256k | Alibaba | 20* | -- | 103 | 2.13 | 7.00 | |
o3 | 200k | OpenAI | 20* | -- | 144 | 5.68 | 9.14 | |
Qwen3.6 27B (Non-reasoning) | 262k | Alibaba | 20* | -- | 58 | 3.63 | 12.19 | |
K-EXAONE 2.0 | 262k |  LG AI Research | 20* | -- | -- | -- | -- | |
LongCat 2.0 | 1M |  LongCat | 20 | $0.06 | -- | -- | -- | |
Step 3.7 Flash | 262k |  StepFun | 19* | -- | 114 | 2.67 | 24.66 | |
Qwen3.5 397B A17B | 262k | Alibaba | 19 | $0.47 | 85 | 2.24 | 45.67 | |
DeepSeek V4 Flash (Non-reasoning) | 1M | DeepSeek | 19* | -- | -- | -- | -- | |
Qwen3.6 35B A3B | 262k | Alibaba | 19 | $0.48 | 123 | 2.07 | 50.02 | |
JT-35B-Flash | 256k | China Mobile | 19* | -- | -- | -- | -- | |
MiMo-V2.5-Pro (Non-reasoning) | 1M | Xiaomi | 18* | -- | 38 | 6.56 | 19.83 | |
Muse Glimmer (high) | 131k | Meta | 18 | $0.06 | 90 | 1.06 | 28.73 | |
Qwen3.5 122B A10B (Non-reasoning) | 262k | Alibaba | 18* | -- | 139 | 2.41 | 6.01 | |
Claude 4.5 Haiku | 200k | Anthropic | 18 | $0.21 | 85 | 17.16 | 23.06 | |
Ring-2.6-1T | 262k | InclusionAI | 17 | $0.29 | 126 | 3.86 | 23.67 | |
Doubao Seed Code | 256k |  ByteDance Seed | 17* | -- | -- | -- | -- | |
GPT-5.6 Luna (Non-reasoning) | 1M | OpenAI | 17* | -- | 113 | 0.87 | 5.32 | |
Gemma 4 26B A4B | 256k | Google | 17* | -- | -- | -- | -- | |
Qwen3.5 122B A10B | 262k | Alibaba | 16 | $0.32 | 127 | 2.34 | 21.97 | |
K2 Horizon 3.7B | 524k | MBZUAI Institute of Foundation Models | 16 | -- | -- | -- | -- | |
MiMo-V2-Flash (Non-reasoning) | 256k | Xiaomi | 16* | -- | -- | -- | -- | |
Gemma 4 31B | 256k | Google | 15 | $0.00 | 35 | 1.01 | 65.04 | |
Claude 4.5 Haiku (Non-reasoning) | 200k | Anthropic | 15* | -- | 81 | 0.70 | 6.89 | |
Qwen3.6 35B A3B (Non-reasoning) | 262k | Alibaba | 15* | -- | 134 | 2.06 | 5.78 | |
Qwen3.5 35B A3B (Non-reasoning) | 262k | Alibaba | 15* | -- | 165 | 2.08 | 5.11 | |
Mistral Medium 3.5 | 256k |  Mistral | 15 | $0.44 | 147 | 2.24 | 19.28 | |
Granite 4.2 30B | 131k | IBM | 15* | -- | 77 | 0.80 | 33.16 | |
Grok 4.3 (Non-reasoning) | 1M | SpaceXAI | 15 | $0.14 | 115 | 0.74 | 5.07 | |
K-EXAONE | 256k | LG AI Research | 14* | -- | -- | -- | -- | |
ERNIE 5.0 Thinking Preview | 128k |  Baidu | 14* | -- | -- | -- | -- | |
Gemma 4 12B | 256k | Google | 14* | -- | 108 | 2.43 | 25.64 | |
Nova 2.0 Pro Preview (medium) | 256k |  Amazon | 14* | -- | 112 | 12.82 | 35.14 | |
Command A+ | 192k | Cohere | 14 | $0.00 | 240 | 0.38 | 10.81 | |
Gemma 4 31B (Non-reasoning) | 256k | Google | 14* | -- | 48 | 1.89 | 12.36 | |
Qwen3.5 9B | 262k | Alibaba | 14* | -- | 88 | 1.93 | 30.20 | |
Nova 2.0 Omni (medium) | 1M | Amazon | 14* | -- | -- | -- | -- | |
Nemotron 3.5 Lightning | 1M | NVIDIA | 14 | $0.09 | 303 | 0.56 | 8.81 | |
Nemotron 3 Super | 1M | NVIDIA | 14 | $1.06 | 173 | 2.39 | 16.83 | |
Apriel-v1.6-15B-Thinker | 128k |  ServiceNow | 13* | -- | -- | -- | -- | |
Nova 2.0 Lite (high) | 1M | Amazon | 13* | -- | 173 | 14.07 | 28.50 | |
Qwen3.5 9B (Non-reasoning) | 262k | Alibaba | 13* | -- | 92 | 0.78 | 6.22 | |
EXAONE 4.5 33B | 262k | LG AI Research | 13* | -- | -- | -- | -- | |
MiniCPM5-2B | 131k |  OpenBMB | 13 | -- | -- | -- | -- | |
Gemma 4 26B A4B (Non-reasoning) | 256k | Google | 13* | -- | 51 | 1.38 | 11.12 | |
Qwen3.5 4B | 262k | Alibaba | 13* | -- | 29 | 0.70 | 85.88 | |
North Mini Code | 256k | Cohere | 13* | -- | 84 | 0.41 | 30.23 | |
Nova 2.0 Pro Preview (low) | 256k | Amazon | 13* | -- | 113 | 10.88 | 32.97 | |
Nova 2.0 Lite (medium) | 1M | Amazon | 12* | -- | 166 | 10.15 | 25.18 | |
Qwen3.5 Omni Flash | 256k | Alibaba | 12* | -- | 225 | 1.90 | 4.12 | |
gpt-oss-120b (high) | 131k | OpenAI | 12 | $0.11 | 229 | 0.84 | 11.75 | |
JT-MINI | 128k | China Mobile | 12* | -- | -- | -- | -- | |
Ling 3.0 Tiny | 262k | InclusionAI | 12 | $0.00 | 171 | 2.63 | 17.27 | |
Magistral Medium 1.2 | 128k | Mistral | 12* | -- | -- | -- | -- | |
Granite 4.2 8B | 131k | IBM | 12 | $0.02 | 70 | 0.63 | 36.31 | |
Nova 2.0 Lite (low) | 1M | Amazon | 12* | -- | 167 | 8.33 | 23.29 | |
HyperNova 60B 2605 (high) | 131k | Multiverse Computing | 12* | -- | 348 | 0.60 | 7.78 | |
Nemotron Cascade 2 30B A3B | 1M | NVIDIA | 12* | -- | -- | -- | -- | |
Mercury 2 | 128k | Inception | 12 | $0.07 | 801 | 4.46 | 5.09 | |
K2 Think V2 | 262k | MBZUAI Institute of Foundation Models | 11* | -- | -- | -- | -- | |
LongCat Flash Lite | 256k | LongCat | 11* | -- | -- | -- | -- | |
Mistral Small 4 | 256k | Mistral | 11 | $0.05 | 160 | 0.77 | 16.42 | |
HyperCLOVA X SEED Think (32B) | 128k |  Naver | 11* | -- | -- | -- | -- | |
K-EXAONE (Non-reasoning) | 256k | LG AI Research | 11* | -- | -- | -- | -- | |
Qwen3 Next 80B A3B (Reasoning) | 262k | Alibaba | 11* | -- | 198 | 2.37 | 14.99 | |
Nova 2.0 Omni (low) | 1M | Amazon | 11* | -- | -- | -- | -- | |
Mi:dm K 2.5 Pro | 128k |  Korea Telecom | 11* | -- | -- | -- | -- | |
Trinity Large Thinking | 512k |  Arcee AI | 11 | $0.35 | 319 | 1.32 | 9.15 | |
G9v3-3B | 131k | AI9Stars | 11* | -- | -- | -- | -- | |
Qwen3.5 4B (Non-reasoning) | 262k | Alibaba | 11* | -- | 20 | 0.74 | 25.43 | |
INTELLECT-3 | 131k |  Prime Intellect | 11* | -- | -- | -- | -- | |
Solar Open 100B | 128k | Upstage | 10* | -- | -- | -- | -- | |
Nemotron 3 Nano Omni 30B A3B | 256k | NVIDIA | 10* | -- | -- | -- | -- | |
gpt-oss-120b (low) | 131k | OpenAI | 10* | -- | 224 | 0.83 | 11.99 | |
Qwen3 Coder Next | 256k | Alibaba | 10 | $0.55 | 108 | 1.16 | 5.77 | |
Nova 2.0 Pro Preview (Non-reasoning) | 256k | Amazon | 10* | -- | 113 | 1.04 | 5.47 | |
gpt-oss-20b (low) | 131k | OpenAI | 10* | -- | 264 | 0.86 | 10.32 | |
K2-V2 (high) | 512k | MBZUAI Institute of Foundation Models | 10* | -- | -- | -- | -- | |
Mistral Large 3 | 256k | Mistral | 10 | $0.10 | 74 | 1.06 | 7.81 | |
Qwen3 Next 80B A3B | 262k | Alibaba | 10* | -- | 181 | 2.24 | 5.01 | |
DiffusionGemma 26B A4B | 256k | Google | 10* | -- | -- | -- | -- | |
Devstral 2 | 256k | Mistral | 9 | $0.00 | 138 | 2.25 | 5.88 | |
Gemma 4 12B (Non-reasoning) | 262k | Google | 9* | -- | 103 | 2.34 | 7.17 | |
Llama 4 Maverick | 1M | Meta | 9 | $0.04 | 97 | 0.86 | 6.00 | |
Motif-2-12.7B | 128k | Motif Technologies | 9* | -- | -- | -- | -- | |
Nova Premier | 1M | Amazon | 9* | -- | 32 | 2.82 | 18.24 | |
Granite 4.2 3B | 131k | IBM | 9 | $0.01 | 218 | 0.45 | 11.93 | |
gpt-oss-20b (high) | 131k | OpenAI | 9 | $0.01 | 225 | 0.73 | 11.84 | |
K2-V2 (medium) | 512k | MBZUAI Institute of Foundation Models | 9* | -- | -- | -- | -- | |
Llama Nemotron Super 49B v1.5 | 128k | NVIDIA | 9* | -- | 95 | 4.25 | 30.66 | |
Mistral Small 4 (Non-reasoning) | 256k | Mistral | 9* | -- | 152 | 0.75 | 4.03 | |
Tri-21B-Think | 32k |  Trillion Labs | 9* | -- | -- | -- | -- | |
Gemma 4 E4B | 128k | Google | 9* | -- | 45 | 0.81 | 56.70 | |
Nemotron 3 Nano | 1M | NVIDIA | 9 | $0.02 | 223 | 1.18 | 12.38 | |
MiniCPM5-1B | 128k | OpenBMB | 9* | -- | -- | -- | -- | |
Sarvam 105B (high) | 128k |  Sarvam | 9* | -- | -- | -- | -- | |
Nova 2.0 Lite (Non-reasoning) | 1M | Amazon | 9* | -- | 155 | 1.07 | 4.31 | |
MiniCPM5-1B (Non-reasoning) | 128k | OpenBMB | 9* | -- | -- | -- | -- | |
Magistral Small 1.2 | 128k | Mistral | 9* | -- | -- | -- | -- | |
Nanbeige4.1-3B | 256k |  Nanbeige | 8* | -- | -- | -- | -- | |
LFM2.5-2.6B | 128k |  Liquid AI | 8* | -- | -- | -- | -- | |
Devstral Small 2 | 256k | Mistral | 8 | $0.00 | 139 | 2.23 | 5.82 | |
EXAONE 4.0 32B | 131k | LG AI Research | 8* | -- | -- | -- | -- | |
Nova 2.0 Omni (Non-reasoning) | 1M | Amazon | 8* | -- | -- | -- | -- | |
Hermes 4 70B | 128k |  Nous Research | 8* | -- | -- | -- | -- | |
Falcon-H1R-7B | 256k |  TII UAE | 8* | -- | -- | -- | -- | |
Solar Pro 3 | 128k | Upstage | 8 | $0.08 | 147 | 2.37 | 19.35 | |
Gemma 4 E2B | 128k | Google | 8* | -- | -- | -- | -- | |
Qwen3 Omni 30B A3B (Reasoning) | 65.5k | Alibaba | 8* | -- | 100 | 1.91 | 26.85 | |
Step3 VL 10B | 65.5k | StepFun | 8* | -- | -- | -- | -- | |
Llama 3.3 70B | 128k | Meta | 8* | -- | 87 | 1.66 | 7.39 | |
Llama Nemotron Ultra | 128k | NVIDIA | 8* | -- | -- | -- | -- | |
ERNIE 4.5 300B A47B | 131k | Baidu | 8* | -- | -- | -- | -- | |
Hermes 4 405B | 128k | Nous Research | 7* | -- | 43 | 2.67 | 60.64 | |
NVIDIA Nemotron Nano 12B v2 VL | 128k | NVIDIA | 7* | -- | 148 | 3.29 | 20.22 | |
Gemma 4 E4B (Non-reasoning) | 128k | Google | 7* | -- | 46 | 0.77 | 11.68 | |
Granite 4.1 30B | 131k | IBM | 7* | -- | -- | -- | -- | |
NVIDIA Nemotron Nano 9B V2 | 131k | NVIDIA | 7* | -- | 95 | 6.83 | 33.20 | |
Hermes 4 405B (Non-reasoning) | 128k | Nous Research | 7* | -- | 43 | 2.41 | 13.96 | |
Nemotron 3 Nano 4B | 262k | NVIDIA | 7* | -- | -- | -- | -- | |
Llama Nemotron Super 49B v1.5 (Non-reasoning) | 128k | NVIDIA | 7* | -- | 161 | 3.00 | 6.11 | |
K2-V2 (low) | 512k | MBZUAI Institute of Foundation Models | 7* | -- | -- | -- | -- | |
Kimi Linear 48B A3B Instruct | 1M | Kimi | 7* | -- | -- | -- | -- | |
Llama 3.1 405B | 128k | Meta | 7* | -- | -- | -- | -- | |
LFM2.5-8B-A1B | 32.8k | Liquid AI | 7* | -- | -- | -- | -- | |
Ring-flash-2.0 | 128k | InclusionAI | 7* | -- | -- | -- | -- | |
Olmo 3.1 32B Think | 65.5k |  Allen Institute for AI | 7* | -- | -- | -- | -- | |
Command A | 256k | Cohere | 7* | -- | 52 | 1.79 | 11.34 | |
Qwen3.5 2B | 262k | Alibaba | 7* | -- | -- | -- | -- | |
Llama 3.1 Nemotron 70B | 128k | NVIDIA | 7* | -- | 126 | 3.06 | 7.02 | |
Nemotron 3 Nano (Non-reasoning) | 1M | NVIDIA | 7* | -- | 174 | 0.97 | 3.85 | |
NVIDIA Nemotron Nano 9B V2 (Non-reasoning) | 131k | NVIDIA | 7* | -- | 159 | 1.61 | 4.76 | |
Hermes 4 70B (Non-reasoning) | 128k | Nous Research | 7* | -- | -- | -- | -- | |
Granite 4.1 8B | 131k | IBM | 7* | -- | 126 | 0.74 | 4.72 | |
Sarvam 30B (high) | 65.5k | Sarvam | 7* | -- | -- | -- | -- | |
Olmo 3.1 32B Instruct | 65.5k | Allen Institute for AI | 6* | -- | -- | -- | -- | |
Gemma 4 E2B (Non-reasoning) | 128k | Google | 6* | -- | -- | -- | -- | |
Llama 4 Scout | 10M | Meta | 6 | $0.12 | 118 | 0.87 | 5.11 | |
R1 1776 | 128k |  Perplexity | 6* | -- | -- | -- | -- | |
Llama 3.2 90B (Vision) | 128k | Meta | 6* | -- | -- | -- | -- | |
Celeris-1 | 131k | Celeris | 6 | $0.05 | 1,418 | 0.60 | 0.95 | |
Phi-4 Mini | 128k |  Microsoft | 6* | -- | 45 | 0.86 | 11.94 | |
EXAONE 4.0 32B (Non-reasoning) | 131k | LG AI Research | 6* | -- | -- | -- | -- | |
Qwen3.5 2B (Non-reasoning) | 262k | Alibaba | 6* | -- | -- | -- | -- | |
Qwen3.5 0.8B | 262k | Alibaba | 6* | -- | -- | -- | -- | |
DeepHermes 3 - Mistral 24B | 32k | Nous Research | 6* | -- | -- | -- | -- | |
Jamba 1.7 Large | 256k |  AI21 Labs | 6* | -- | -- | -- | -- | |
Granite 4.0 H Small | 128k | IBM | 6* | -- | 16 | 29.14 | 60.37 | |
Ministral 3 14B | 256k | Mistral | 6 | $0.14 | 82 | 0.93 | 7.03 | |
Qwen3 Omni 30B A3B | 65.5k | Alibaba | 6* | -- | 96 | 1.91 | 7.09 | |
LFM2 24B A2B | 32.8k | Liquid AI | 6* | -- | -- | -- | -- | |
Phi-4 | 16k | Microsoft | 6* | -- | 40 | 2.54 | 14.91 | |
Nova Micro | 130k | Amazon | 6* | -- | 269 | 0.89 | 2.74 | |
Granite 4.1 3B | 131k | IBM | 6* | -- | -- | -- | -- | |
NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning) | 128k | NVIDIA | 6* | -- | 198 | 1.18 | 3.71 | |
Phi-4 Multimodal | 128k | Microsoft | 6* | -- | 17 | 0.87 | 30.19 | |
MiniCPM-V 4.6 1.3B | 262k | OpenBMB | 6* | -- | -- | -- | -- | |
Jamba Reasoning 3B | 262k | AI21 Labs | 6* | -- | -- | -- | -- | |
Reka Flash 3 | 128k |  Reka AI | 6* | -- | -- | -- | -- | |
Olmo 3 7B Think | 65.5k | Allen Institute for AI | 6* | -- | -- | -- | -- | |
Molmo 7B-D | 4.1k | Allen Institute for AI | 6* | -- | -- | -- | -- | |
Ministral 3 8B | 256k | Mistral | 5 | $0.07 | 89 | 0.73 | 6.34 | |
Llama 3.2 11B (Vision) | 128k | Meta | 5* | -- | 19 | 1.27 | 27.17 | |
Qwen3.5 0.8B (Non-reasoning) | 262k | Alibaba | 5* | -- | -- | -- | -- | |
Exaone 4.0 1.2B | 64k | LG AI Research | 5* | -- | -- | -- | -- | |
Olmo 3 7B | 65.5k | Allen Institute for AI | 5* | -- | -- | -- | -- | |
Exaone 4.0 1.2B (Non-reasoning) | 64k | LG AI Research | 5* | -- | -- | -- | -- | |
LFM2.5-1.2B-Thinking | 32k | Liquid AI | 5* | -- | -- | -- | -- | |
Jamba 1.7 Mini | 258k | AI21 Labs | 5* | -- | -- | -- | -- | |
LFM2 2.6B | 32.8k | Liquid AI | 5* | -- | -- | -- | -- | |
LFM2.5-1.2B-Instruct | 32k | Liquid AI | 5* | -- | -- | -- | -- | |
Granite 4.0 H 1B | 128k | IBM | 5* | -- | -- | -- | -- | |
Gemma 3 270M | 32k | Google | 5* | -- | -- | -- | -- | |
Apertus 70B Instruct | 65.5k |  Swiss AI Initiative | 5* | -- | -- | -- | -- | |
Granite 4.0 Micro | 128k | IBM | 5* | -- | -- | -- | -- | |
DeepHermes 3 - Llama-3.1 8B | 128k | Nous Research | 5* | -- | -- | -- | -- | |
Granite 4.0 1B | 128k | IBM | 5* | -- | -- | -- | -- | |
Molmo2-8B | 36.9k | Allen Institute for AI | 5* | -- | -- | -- | -- | |
LFM2 8B A1B | 32.8k | Liquid AI | 5* | -- | -- | -- | -- | |
Ministral 3 3B | 256k | Mistral | 5 | $0.04 | 213 | 0.61 | 2.96 | |
LFM2.5-VL-1.6B | 32k | Liquid AI | 5* | -- | -- | -- | -- | |
Granite 4.0 350M | 32.8k | IBM | 5* | -- | -- | -- | -- | |
Tiny Aya Global | 8.19k | Cohere | 5* | -- | -- | -- | -- | |
Apertus 8B Instruct | 65.5k | Swiss AI Initiative | 5* | -- | -- | -- | -- | |
Granite 4.0 H 350M | 32.8k | IBM | 5* | -- | -- | -- | -- | |
K2 Horizon 0.9B | 131k | MBZUAI Institute of Foundation Models | 3 | -- | -- | -- | -- | |
EXAONE 4.5 33B (Non-reasoning) | 262k | LG AI Research | -- | -- | -- | -- | -- | |
Gemini 3 Deep Think | 128k | Google | -- | -- | -- | -- | -- | |
GPT-5.5 Pro (xhigh) | 922k | OpenAI | -- | -- | -- | -- | -- | |
Cogito v2.1 | 128k |  Deep Cogito | -- | -- | -- | -- | -- | |
Principais definições
Perguntas frequentes
Atualmente, Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) ocupa o 1º lugar no ranking de LLMs da Artificial Analysis, com 53 pontos no Artificial Analysis Intelligence Index, entre 155 modelos classificados.
Os melhores modelos segundo o Artificial Analysis Intelligence Index são: 1. Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) (53), 2. Claude Fable 5.1 (Adaptive Reasoning, Xhigh Effort, Default Fallback) (53), 3. GPT-6 Astra (max) (53), 4. GPT-6 Astra (xhigh) (53), 5. Claude Fable 5.1 (Adaptive Reasoning, High Effort, Default Fallback) (51).
Celeris-1 é o mais rápido, com 1,418.0 tokens por segundo, seguido por Mercury 2 (801.3 t/s) e Gemini 2.5 Flash-Lite (Reasoning) (423.6 t/s).
Granite 4.2 3B tem o menor custo por tarefa do Artificial Analysis Intelligence Index, de $0.01, seguido por GPT-5.6 Luna (low) ($0.01) e gpt-oss-20b (high) ($0.01).
GLM-5.3 (max) é o modelo de pesos abertos mais bem classificado, com 45 pontos no Artificial Analysis Intelligence Index. O ranking tem 78 modelos de pesos abertos entre 155 modelos no total.
Os melhores modelos de pesos abertos segundo o Artificial Analysis Intelligence Index são: 1. GLM-5.3 (max) (45), 2. Kimi K3 (max) (44), 3. GLM-5.3-Flash (42).
Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) lidera entre 138 modelos de raciocínio, com 53 pontos no Artificial Analysis Intelligence Index. Os modelos de raciocínio usam um processo de reflexão prolongado para resolver problemas complexos antes de responder.
O ranking inclui filtros para restringir os resultados por tipo de modelo (com ou sem raciocínio), abertura (pesos abertos ou proprietário) e outros critérios. Você também pode ajustar as opções de prompt para ver como o desempenho varia conforme a extensão da entrada.
Clique no nome de qualquer modelo no ranking para acessar sua página de comparação, com gráficos detalhados de inteligência, preços, velocidade, latência e outras métricas. Você também pode comparar os provedores de API de cada modelo. Ver todos os modelos