Ranking de modelos de lenguaje (LLM) - Comparación de modelos de IA de OpenAI, Anthropic, Google, SpaceXAI y otros
Comparación y ranking del rendimiento de más de 250 modelos de IA (LLM) en métricas clave como inteligencia, precio, rendimiento y velocidad (velocidad de salida - tokens por segundo y latencia - TTFT), ventana de contexto y otras.
Para más detalles, incluso sobre nuestra metodología, consulta nuestras preguntas frecuentes.
Inteligencia
UpdatedClaude Fable 5.1 (max with fallback) y GPT-6 Astra (max) son los modelos con mayor inteligencia, seguidos por GPT-6 Astra (xhigh) y Claude Opus 5 (max).
Velocidad de salida
Celeris-1 y Mercury 2 son los modelos más rápidos, seguidos por Gemini 2.5 Flash-Lite y Gemini 3.5 Flash-Lite.
Latencia
Gemini 2.5 Flash-Lite (Non-reasoning) y Command A+ son los modelos con menor latencia, seguidos por Granite 4.2 3B y North Mini Code.
Costo por tarea
Granite 4.2 3B y Granite 4.2 8B tienen el menor costo por tarea, seguidos por Ling 3.0 Flash y MiMo-V2.5-Pro.
Ventana de contexto
Llama 4 Scout y Grok 4.20 0309 admiten las ventanas de contexto más grandes, seguidos por Gemini 1.5 Pro (May) y Grok 4.1 Fast.
Análisis adicional | ||||||||
|---|---|---|---|---|---|---|---|---|
Claude Fable 5.1 (max with fallback) | 1M | Anthropic | 57 | $6.12 | 67 | 266.51 | 273.96 | |
GPT-6 Astra (max) | 1M | OpenAI | 55 | $2.57 | 62 | 250.18 | 258.19 | |
GPT-6 Astra (xhigh) | 1M | OpenAI | 54 | $1.85 | 57 | 119.44 | 128.25 | |
Claude Opus 5 (max) | 1M | Anthropic | 54 | $4.21 | 58 | 73.62 | 82.31 | |
Claude Fable 5.1 (xhigh with fallback) | 1M | Anthropic | 54* | -- | 69 | 102.62 | 109.87 | |
Claude Opus 5 (xhigh) | 1M | Anthropic | 53 | $3.36 | 56 | 33.78 | 42.75 | |
GPT-6 Astra (high) | 1M | OpenAI | 53 | $1.41 | 61 | 32.71 | 40.94 | |
Claude Fable 5 (with fallback) | 1M | Anthropic | 53 | $5.62 | 70 | 105.76 | 112.95 | |
Muse Spark 1.3 (max) | 1M | Meta | 53 | $0.96 | 190 | 18.69 | 31.84 | |
GPT-6 Astra (medium) | 1M | OpenAI | 52 | $1.16 | 58 | 8.07 | 16.73 | |
Claude Opus 5 (high) | 1M | Anthropic | 52 | $2.44 | 55 | 19.49 | 28.66 | |
Claude Fable 5.1 (high with fallback) | 1M | Anthropic | 52* | -- | 56 | 28.88 | 37.79 | |
Muse Spark 1.3 (xhigh) | 1M | Meta | 52 | $0.84 | 120 | 42.49 | 63.25 | |
GPT-5.6 Sol (max) | 1M | OpenAI | 51 | $1.25 | 85 | 107.18 | 113.03 | |
Grok 4.6 (high) | 500k | SpaceXAI | 51 | $1.25 | 64 | 52.03 | 59.85 | |
Kimi K3 (max) | 1.05M |  Kimi | 50 | $1.58 | 40 | 4.29 | 66.37 | |
Claude Fable 5.1 (medium with fallback) | 1M | Anthropic | 50* | -- | 57 | 15.81 | 24.58 | |
GPT-5.6 Sol (xhigh) | 1M | OpenAI | 50 | $0.89 | 81 | 44.70 | 50.87 | |
Claude Opus 5 (medium) | 1M | Anthropic | 50 | $1.36 | 53 | 5.78 | 15.22 | |
Grok 4.6 (xhigh) | 500k | SpaceXAI | 49* | -- | 64 | 55.14 | 63.00 | |
GPT-6 Astra (low) | 1M | OpenAI | 49 | $0.63 | 59 | 1.95 | 10.38 | |
GLM-5.3 (max) | 1M |  Z AI | 49 | $1.26 | 84 | 2.14 | 31.84 | |
Grok 4.6 (medium) | 500k | SpaceXAI | 48* | -- | 62 | 39.09 | 47.14 | |
GPT-5.6 Sol (high) | 1M | OpenAI | 48 | $0.61 | 73 | 13.81 | 20.64 | |
GPT-6 Astra (Non-reasoning) | 1M | OpenAI | 48 | $1.42 | -- | -- | -- | |
Claude Fable 5.1 (low with fallback) | 1M | Anthropic | 48* | -- | 56 | 6.58 | 15.54 | |
Gemini 3.8 Flash (high) | 1M | Google | 47 | $0.74 | -- | -- | -- | |
Qwen3.8 Max | 1M |  Alibaba | 47 | $1.19 | 40 | 2.35 | 65.28 | |
Muse Spark 1.2 (xhigh) | 1.05M | Meta | 47 | $0.55 | 260 | 15.13 | 24.75 | |
Gemini 3.8 Flash (medium) | 1M | Google | 47* | -- | -- | -- | -- | |
GPT-5.6 Terra (max) | 1M | OpenAI | 47 | $0.81 | 110 | 165.94 | 170.49 | |
Qwen3.8 2.4T A95B | 984k | Alibaba | 47 | $1.10 | 40 | 2.53 | 64.82 | |
GLM-5.3-Flash | 1M | Z AI | 46 | $0.18 | 47 | 1.63 | 55.03 | |
GPT-5.6 Sol (medium) | 1M | OpenAI | 46 | $0.37 | 72 | 4.92 | 11.90 | |
Qwen3.8-Flash-Next | 256k | Alibaba | 46* | -- | 66 | 2.68 | 40.58 | |
Grok 4.5 (high) | 500k | SpaceXAI | 45 | $0.43 | 58 | 8.57 | 17.25 | |
Gemini 3.7 Flash (high) | 1M | Google | 45 | $0.55 | 309 | 10.83 | 12.45 | |
Claude Sonnet 5 (max) | 1M | Anthropic | 45 | $3.31 | 75 | 190.27 | 196.98 | |
GPT-5.6 Terra (xhigh) | 1M | OpenAI | 44 | $0.50 | 86 | 27.98 | 33.81 | |
Claude Opus 5 (low) | 1M | Anthropic | 44 | $0.64 | 54 | 2.94 | 12.22 | |
GPT-5.6 Luna (max) | 1M | OpenAI | 43 | $0.10 | 133 | 124.70 | 128.46 | |
Gemini 3.7 Flash (medium) | 1M | Google | 43* | -- | 299 | 5.29 | 6.96 | |
GLM-5.2 (max) | 1M | Z AI | 43* | -- | 71 | 1.61 | 36.63 | |
Gemini 3.8 Flash (low) | 1M | Google | 42* | -- | -- | -- | -- | |
DeepSeek V4 Pro 0813 (max) | 1M |  DeepSeek | 42 | $0.33 | 66 | 1.70 | 39.63 | |
Grok 4.6 (low) | 500k | SpaceXAI | 42* | -- | 55 | 8.05 | 17.09 | |
GPT-5.6 Luna (xhigh) | 1M | OpenAI | 42 | $0.06 | 117 | 65.08 | 69.36 | |
DeepSeek V4 Flash Vision (max) | 1M | DeepSeek | 42* | -- | 120 | 1.08 | 21.89 | |
Qwen3.8 27B (xhigh) | 256k | Alibaba | 41 | $0.57 | 44 | 4.00 | 61.04 | |
GPT-5.6 Terra (high) | 1M | OpenAI | 41 | $0.30 | 87 | 4.83 | 10.60 | |
Gemini 3.7 Flash (low) | 1M | Google | 41* | -- | 286 | 0.90 | 2.65 | |
GPT-5.6 Sol (low) | 1M | OpenAI | 41 | $0.23 | 67 | 3.35 | 10.78 | |
DeepSeek V4 Flash 0731 (max) | 1M | DeepSeek | 41 | $0.14 | 134 | 1.05 | 19.68 | |
Gemini 3.6 Flash | 1M | Google | 40 | $0.36 | 215 | 15.82 | 18.14 | |
Agnes 2.5 Pro Beta | 1M |  Sapiens AI | 40* | -- | -- | -- | -- | |
Kimi K3 (low) | 1.05M | Kimi | 39* | -- | 39 | 4.53 | 69.10 | |
Gemini 3.5 Flash (medium) | 1M | Google | 38* | -- | 223 | 14.22 | 16.46 | |
Motif 3 | 262k |  Motif Technologies | 38* | -- | -- | -- | -- | |
K2 Horizon 375B A23B | 524k |  MBZUAI Institute of Foundation Models | 38 | -- | -- | -- | -- | |
GPT-5.6 Luna (high) | 1M | OpenAI | 37* | -- | 123 | 19.71 | 23.78 | |
GPT-5.6 Terra (medium) | 1M | OpenAI | 37* | -- | 83 | 1.98 | 7.97 | |
GPT-5.3 Codex (xhigh) | 400k | OpenAI | 37* | -- | 134 | 60.05 | 63.79 | |
Motif 3 (Beta) | 262k | Motif Technologies | 37* | -- | -- | -- | -- | |
Gemini 3.1 Pro Preview | 1M | Google | 37 | $0.34 | 115 | 30.82 | 35.17 | |
DeepSeek V4 Pro (max) | 1M | DeepSeek | 36* | -- | 62 | 1.64 | 80.61 | |
MiniMax-M3 | 1M |  MiniMax | 36 | $0.23 | 86 | 1.53 | 30.51 | |
Qwen3.8 27B (medium) | 256k | Alibaba | 35* | -- | 52 | 3.89 | 52.23 | |
Apodex 1.1 | 256k |  Apodex | 35* | -- | -- | -- | -- | |
DeepSeek V4 Pro (high) | 1M | DeepSeek | 35* | -- | 64 | 1.60 | 40.69 | |
Quasar 438B (max) | 1M |  Multiverse Computing | 34* | -- | 182 | 0.97 | 14.74 | |
Kimi K2.7 Code | 256k | Kimi | 34* | -- | 45 | 2.93 | 62.97 | |
Qwen3.8 27B (low) | 256k | Alibaba | 34* | -- | 52 | 3.93 | 51.81 | |
Claude Sonnet 5 (Non-reasoning) | 1M | Anthropic | 33* | -- | 63 | 1.28 | 9.19 | |
Hy3 | 256k |  Tencent | 33* | -- | 95 | 2.78 | 28.98 | |
GPT-5.6 Sol (Non-reasoning) | 1M | OpenAI | 33* | -- | 67 | 0.95 | 8.41 | |
Nex-N2-Pro | 262k |  Nex AGI | 33* | -- | 129 | 1.73 | 21.18 | |
Solar Pro 4 | 512k |  Upstage | 33* | -- | 70 | 1.98 | 37.65 | |
MiMo-V2.5-Pro | 1M | Xiaomi | 33 | $0.04 | 35 | 5.15 | 75.94 | |
GPT-5.6 Terra (low) | 1M | OpenAI | 32* | -- | 87 | 1.78 | 7.50 | |
Inkling Small | 1M |  Thinking Machines | 32* | -- | 112 | 2.53 | 24.95 | |
Inkling | 1M | Thinking Machines | 32 | $0.44 | 73 | 2.73 | 37.15 | |
JT-4.1 Flash 236B A21B | 256k |  China Mobile | 32* | -- | -- | -- | -- | |
Agnes 2.5 Pro Alpha | 1M | Sapiens AI | 31* | -- | 200 | 2.78 | 15.29 | |
Qwen3.7 Plus | 1M | Alibaba | 31* | -- | 56 | 2.16 | 47.14 | |
GPT-5.6 Luna (medium) | 1M | OpenAI | 30* | -- | 107 | 3.01 | 7.68 | |
Nemotron 3 Ultra | 262k |  NVIDIA | 30* | -- | 169 | 2.61 | 18.99 | |
MiMo-V2.5 | 1M | Xiaomi | 30* | -- | 52 | 3.81 | 51.58 | |
MiMo-V2-Omni-0327 | 256k | Xiaomi | 29* | -- | -- | -- | -- | |
Grok 4.3 (medium) | 1M | SpaceXAI | 29* | -- | 120 | 13.47 | 17.64 | |
Solar Open2 250B | 1.05M | Upstage | 29* | -- | -- | -- | -- | |
Qwen3.6 27B | 262k | Alibaba | 29* | -- | 55 | 3.78 | 116.09 | |
Grok 4.3 (low) | 1M | SpaceXAI | 29* | -- | 118 | 5.46 | 9.71 | |
MiMo-V2-Omni | 256k | Xiaomi | 28* | -- | -- | -- | -- | |
Gemini 3.5 Flash (minimal) | 1M | Google | 28* | -- | 189 | 0.97 | 3.61 | |
Gemini 3.5 Flash-Lite | 1M | Google | 28 | $0.08 | 371 | 7.47 | 8.82 | |
Claude Sonnet 4.6 (Non-reasoning, low) | 1M | Anthropic | 27* | -- | 41 | 1.89 | 14.07 | |
Ling 3.0 Flash | 262k | InclusionAI | 27 | $0.03 | 328 | 2.59 | 10.20 | |
A.X-K2 | 262k |  SK Telecom | 27* | -- | -- | -- | -- | |
GLM-5.2 (Non-reasoning) | 1M | Z AI | 26* | -- | 110 | 1.91 | 6.44 | |
Qwen3.8 27B | 256k | Alibaba | 26* | -- | 50 | 3.87 | 13.79 | |
MiMo-V2-Flash (Feb 2026) | 256k | Xiaomi | 26* | -- | -- | -- | -- | |
GPT-5.6 Terra (Non-reasoning) | 1M | OpenAI | 26* | -- | 82 | 0.81 | 6.91 | |
Qwen3.6 35B A3B | 262k | Alibaba | 26* | -- | 129 | 2.03 | 47.85 | |
Qwen3.5 397B A17B | 262k | Alibaba | 26* | -- | 91 | 2.19 | 42.64 | |
LongCat 2.0 | 1M |  LongCat | 26* | -- | 50 | 3.36 | 53.35 | |
GPT-5.6 Luna (low) | 1M | OpenAI | 26* | -- | 110 | 1.72 | 6.26 | |
G9v3-39A5B | 131k |  AI9Stars | 26* | -- | -- | -- | -- | |
KAT-Coder-Pro V2 | 256k |  KwaiKAT | 26* | -- | -- | -- | -- | |
Qwen3.5 397B A17B (Non-reasoning) | 262k | Alibaba | 25* | -- | 87 | 2.24 | 8.00 | |
Qwen3.5 122B A10B | 262k | Alibaba | 25* | -- | 133 | 2.40 | 21.20 | |
DeepSeek V4 Pro (Non-reasoning) | 1M | DeepSeek | 25* | -- | 59 | 1.66 | 10.09 | |
Muse Glimmer (high) | 131k | Meta | 24 | $0.04 | 99 | 1.12 | 26.25 | |
Qwen3.5 Omni Plus | 256k | Alibaba | 24* | -- | 124 | 2.18 | 6.20 | |
Ring-2.6-1T | 262k | InclusionAI | 24* | -- | 128 | 3.91 | 23.39 | |
o3 | 200k | OpenAI | 24* | -- | 116 | 6.34 | 10.66 | |
Qwen3.6 27B (Non-reasoning) | 262k | Alibaba | 23* | -- | 58 | 3.61 | 12.27 | |
K-EXAONE 2.0 | 262k |  LG AI Research | 23* | -- | -- | -- | -- | |
Step 3.7 Flash | 262k |  StepFun | 23* | -- | 91 | 2.76 | 30.11 | |
Mistral Medium 3.5 | 256k |  Mistral | 23* | -- | 129 | 2.31 | 21.67 | |
Claude 4.5 Haiku | 200k | Anthropic | 22 | $0.20 | 100 | 20.20 | 25.21 | |
Gemma 4 31B | 256k | Google | 22* | -- | 35 | 1.03 | 64.92 | |
DeepSeek V4 Flash (Non-reasoning) | 1M | DeepSeek | 22* | -- | -- | -- | -- | |
JT-35B-Flash | 256k | China Mobile | 22* | -- | -- | -- | -- | |
GPT-5.5 Instant (June 2026) | 400k | OpenAI | 22* | -- | 134 | 1.07 | 19.77 | |
MiMo-V2.5-Pro (Non-reasoning) | 1M | Xiaomi | 21* | -- | 33 | 4.54 | 19.74 | |
Qwen3.5 122B A10B (Non-reasoning) | 262k | Alibaba | 21* | -- | 145 | 2.32 | 5.78 | |
Doubao Seed Code | 256k |  ByteDance Seed | 20* | -- | -- | -- | -- | |
GPT-5.6 Luna (Non-reasoning) | 1M | OpenAI | 19* | -- | 110 | 0.80 | 5.34 | |
Gemma 4 26B A4B | 256k | Google | 19* | -- | -- | -- | -- | |
Nemotron 3 Super | 1M | NVIDIA | 19* | -- | 142 | 1.81 | 19.46 | |
MiMo-V2-Flash (Non-reasoning) | 256k | Xiaomi | 18* | -- | -- | -- | -- | |
Grok 4.3 (Non-reasoning) | 1M | SpaceXAI | 17* | -- | 105 | 0.76 | 5.52 | |
Claude 4.5 Haiku (Non-reasoning) | 200k | Anthropic | 17* | -- | 83 | 0.77 | 6.80 | |
Qwen3.6 35B A3B (Non-reasoning) | 262k | Alibaba | 17* | -- | 142 | 2.10 | 5.63 | |
Qwen3.5 35B A3B (Non-reasoning) | 262k | Alibaba | 17* | -- | 170 | 2.13 | 5.06 | |
Granite 4.2 30B | 131k | IBM | 17* | -- | 76 | 0.84 | 33.71 | |
Nemotron 3.5 Lightning | 1M | NVIDIA | 16* | -- | 296 | 0.64 | 9.09 | |
K-EXAONE | 256k | LG AI Research | 16* | -- | -- | -- | -- | |
Ling 3.0 Tiny | 262k | InclusionAI | 16 | $0.00 | 155 | 2.51 | 18.62 | |
ERNIE 5.0 Thinking Preview | 128k |  Baidu | 16* | -- | -- | -- | -- | |
Command A+ | 192k | Cohere | 16* | -- | 254 | 0.40 | 10.24 | |
gpt-oss-120b (high) | 131k | OpenAI | 16 | $0.12 | 162 | 0.81 | 16.23 | |
Gemma 4 12B | 256k | Google | 16* | -- | 112 | 2.32 | 24.58 | |
Nova 2.0 Pro Preview (medium) | 256k |  Amazon | 16* | -- | 117 | 15.94 | 37.23 | |
Gemma 4 31B (Non-reasoning) | 256k | Google | 15* | -- | 57 | 1.42 | 10.16 | |
Mercury 2 | 128k | Inception | 15* | -- | 821 | 5.36 | 5.97 | |
Qwen3.5 9B | 262k | Alibaba | 15* | -- | 82 | 1.90 | 32.44 | |
Nova 2.0 Omni (medium) | 1M | Amazon | 15* | -- | -- | -- | -- | |
Apriel-v1.6-15B-Thinker | 128k |  ServiceNow | 14* | -- | -- | -- | -- | |
Qwen3 Coder Next | 256k | Alibaba | 14* | -- | 103 | 1.54 | 6.38 | |
Nova 2.0 Lite (high) | 1M | Amazon | 14* | -- | 163 | 18.79 | 34.11 | |
Qwen3.5 9B (Non-reasoning) | 262k | Alibaba | 14* | -- | 86 | 0.83 | 6.63 | |
EXAONE 4.5 33B | 262k | LG AI Research | 14* | -- | -- | -- | -- | |
Granite 4.2 8B | 131k | IBM | 14 | $0.02 | 143 | 0.62 | 18.06 | |
Gemma 4 26B A4B (Non-reasoning) | 256k | Google | 14* | -- | 77 | 1.34 | 7.85 | |
Qwen3.5 4B | 262k | Alibaba | 14* | -- | 23 | 0.75 | 110.42 | |
North Mini Code | 256k | Cohere | 13* | -- | 56 | 0.43 | 45.45 | |
Nova 2.0 Pro Preview (low) | 256k | Amazon | 13* | -- | 117 | 10.14 | 31.58 | |
Nova 2.0 Lite (medium) | 1M | Amazon | 13* | -- | 155 | 16.53 | 32.64 | |
Qwen3.5 Omni Flash | 256k | Alibaba | 13* | -- | 236 | 1.84 | 3.95 | |
Mistral Small 4 | 256k | Mistral | 13* | -- | 163 | 0.80 | 16.13 | |
Devstral 2 | 256k | Mistral | 13* | -- | 118 | 2.26 | 6.49 | |
JT-MINI | 128k | China Mobile | 12* | -- | -- | -- | -- | |
Trinity Large Thinking | 512k |  Arcee AI | 12* | -- | 322 | 1.19 | 8.95 | |
Magistral Medium 1.2 | 128k | Mistral | 12* | -- | -- | -- | -- | |
Nova 2.0 Lite (low) | 1M | Amazon | 12* | -- | 168 | 8.34 | 23.18 | |
HyperNova 60B 2605 (high) | 131k | Multiverse Computing | 12* | -- | 345 | 0.80 | 8.04 | |
Nemotron Cascade 2 30B A3B | 1M | NVIDIA | 12* | -- | -- | -- | -- | |
K2 Think V2 | 262k | MBZUAI Institute of Foundation Models | 11* | -- | -- | -- | -- | |
LongCat Flash Lite | 256k | LongCat | 11* | -- | -- | -- | -- | |
Devstral Small 2 | 256k | Mistral | 11* | -- | 132 | 2.25 | 6.03 | |
Mistral Large 3 | 256k | Mistral | 11 | $0.07 | 77 | 1.12 | 7.64 | |
HyperCLOVA X SEED Think (32B) | 128k |  Naver | 11* | -- | -- | -- | -- | |
K-EXAONE (Non-reasoning) | 256k | LG AI Research | 11* | -- | -- | -- | -- | |
Granite 4.2 3B | 131k | IBM | 11 | $0.01 | 211 | 0.43 | 12.31 | |
Qwen3 Next 80B A3B (Reasoning) | 262k | Alibaba | 11* | -- | 191 | 2.26 | 15.35 | |
Nova 2.0 Omni (low) | 1M | Amazon | 11* | -- | -- | -- | -- | |
Mi:dm K 2.5 Pro | 128k |  Korea Telecom | 10* | -- | -- | -- | -- | |
G9v3-3B | 131k | AI9Stars | 10* | -- | -- | -- | -- | |
Qwen3.5 4B (Non-reasoning) | 262k | Alibaba | 10* | -- | 18 | 0.76 | 27.85 | |
INTELLECT-3 | 131k |  Prime Intellect | 10* | -- | -- | -- | -- | |
Solar Open 100B | 128k | Upstage | 9* | -- | -- | -- | -- | |
gpt-oss-20b (high) | 131k | OpenAI | 9* | -- | 168 | 0.83 | 15.74 | |
Nemotron 3 Nano Omni 30B A3B | 256k | NVIDIA | 9* | -- | 328 | 1.11 | 8.73 | |
gpt-oss-120b (low) | 131k | OpenAI | 9* | -- | 189 | 0.85 | 14.09 | |
Nemotron 3 Nano | 1M | NVIDIA | 9* | -- | 159 | 1.27 | 17.04 | |
Solar Pro 3 | 128k | Upstage | 9* | -- | 130 | 2.41 | 21.60 | |
Llama 4 Maverick | 1M | Meta | 9* | -- | 82 | 0.91 | 7.00 | |
Nova 2.0 Pro Preview (Non-reasoning) | 256k | Amazon | 8* | -- | 104 | 1.07 | 5.88 | |
gpt-oss-20b (low) | 131k | OpenAI | 8* | -- | 227 | 0.87 | 11.88 | |
K2-V2 (high) | 512k | MBZUAI Institute of Foundation Models | 8* | -- | -- | -- | -- | |
Qwen3 Next 80B A3B | 262k | Alibaba | 8* | -- | 175 | 2.17 | 5.03 | |
DiffusionGemma 26B A4B | 256k | Google | 8* | -- | -- | -- | -- | |
Gemma 4 12B (Non-reasoning) | 262k | Google | 7* | -- | 110 | 2.43 | 6.99 | |
Motif-2-12.7B | 128k | Motif Technologies | 7* | -- | -- | -- | -- | |
Nova Premier | 1M | Amazon | 7* | -- | 31 | 2.85 | 18.91 | |
K2-V2 (medium) | 512k | MBZUAI Institute of Foundation Models | 7* | -- | -- | -- | -- | |
Llama Nemotron Super 49B v1.5 | 128k | NVIDIA | 7* | -- | 18 | 24.52 | 163.86 | |
Mistral Small 4 (Non-reasoning) | 256k | Mistral | 7* | -- | 148 | 0.78 | 4.15 | |
Tri-21B-Think | 32k |  Trillion Labs | 7* | -- | -- | -- | -- | |
Gemma 4 E4B | 128k | Google | 6* | -- | 39 | 0.91 | 64.59 | |
Celeris-1 | 131k | Celeris | 6* | -- | 1,530 | 0.59 | 0.91 | |
MiniCPM5-1B | 128k |  OpenBMB | 6* | -- | -- | -- | -- | |
Sarvam 105B (high) | 128k |  Sarvam | 6* | -- | -- | -- | -- | |
Nova 2.0 Lite (Non-reasoning) | 1M | Amazon | 6* | -- | 153 | 1.09 | 4.35 | |
MiniCPM5-1B (Non-reasoning) | 128k | OpenBMB | 6* | -- | -- | -- | -- | |
Magistral Small 1.2 | 128k | Mistral | 6* | -- | -- | -- | -- | |
Ministral 3 14B | 256k | Mistral | 6* | -- | 88 | 0.96 | 6.65 | |
Nanbeige4.1-3B | 256k |  Nanbeige | 5* | -- | -- | -- | -- | |
LFM2.5-2.6B | 128k |  Liquid AI | 5* | -- | 201 | 1.59 | 14.04 | |
EXAONE 4.0 32B | 131k | LG AI Research | 5* | -- | -- | -- | -- | |
Nova 2.0 Omni (Non-reasoning) | 1M | Amazon | 5* | -- | -- | -- | -- | |
Llama 4 Scout | 10M | Meta | 5* | -- | 87 | 0.83 | 6.59 | |
Hermes 4 70B | 128k |  Nous Research | 4* | -- | 86 | 1.41 | 30.63 | |
Falcon-H1R-7B | 256k |  TII UAE | 4* | -- | -- | -- | -- | |
Gemma 4 E2B | 128k | Google | 4* | -- | -- | -- | -- | |
Qwen3 Omni 30B A3B (Reasoning) | 65.5k | Alibaba | 4* | -- | 100 | 1.91 | 26.80 | |
Step3 VL 10B | 65.5k | StepFun | 4* | -- | -- | -- | -- | |
Llama 3.3 70B | 128k | Meta | 4* | -- | 87 | 1.69 | 7.44 | |
Ministral 3 8B | 256k | Mistral | 3* | -- | 96 | 0.72 | 5.92 | |
Llama Nemotron Ultra | 128k | NVIDIA | 3* | -- | 52 | 2.41 | 50.17 | |
ERNIE 4.5 300B A47B | 131k | Baidu | 3* | -- | -- | -- | -- | |
Hermes 4 405B | 128k | Nous Research | 3* | -- | 42 | 2.47 | 62.23 | |
NVIDIA Nemotron Nano 12B v2 VL | 128k | NVIDIA | 3* | -- | 22 | 15.41 | 129.27 | |
Gemma 4 E4B (Non-reasoning) | 128k | Google | 3* | -- | 41 | 0.84 | 13.06 | |
Granite 4.1 30B | 131k | IBM | 3* | -- | -- | -- | -- | |
NVIDIA Nemotron Nano 9B V2 | 131k | NVIDIA | 3* | -- | 81 | 9.20 | 40.16 | |
Hermes 4 405B (Non-reasoning) | 128k | Nous Research | 3* | -- | 43 | 2.37 | 13.92 | |
Nemotron 3 Nano 4B | 262k | NVIDIA | 3* | -- | -- | -- | -- | |
Llama Nemotron Super 49B v1.5 (Non-reasoning) | 128k | NVIDIA | 3* | -- | 18 | 20.18 | 48.27 | |
K2-V2 (low) | 512k | MBZUAI Institute of Foundation Models | 3* | -- | -- | -- | -- | |
Kimi Linear 48B A3B Instruct | 1M | Kimi | 3* | -- | -- | -- | -- | |
Llama 3.1 405B | 128k | Meta | 3* | -- | -- | -- | -- | |
LFM2.5-8B-A1B | 32.8k | Liquid AI | 3* | -- | 334 | 1.34 | 8.84 | |
Ring-flash-2.0 | 128k | InclusionAI | 3* | -- | -- | -- | -- | |
Olmo 3.1 32B Think | 65.5k |  Allen Institute for AI | 2* | -- | -- | -- | -- | |
Command A | 256k | Cohere | 2* | -- | 57 | 1.60 | 10.43 | |
Qwen3.5 2B | 262k | Alibaba | 2* | -- | -- | -- | -- | |
Llama 3.1 Nemotron 70B | 128k | NVIDIA | 2* | -- | 17 | 15.60 | 45.31 | |
Nemotron 3 Nano (Non-reasoning) | 1M | NVIDIA | 2* | -- | 157 | 0.89 | 4.07 | |
NVIDIA Nemotron Nano 9B V2 (Non-reasoning) | 131k | NVIDIA | 2* | -- | 134 | 2.28 | 6.01 | |
Ministral 3 3B | 256k | Mistral | 2* | -- | 216 | 0.61 | 2.92 | |
Hermes 4 70B (Non-reasoning) | 128k | Nous Research | 1* | -- | 84 | 1.44 | 7.39 | |
Granite 4.1 8B | 131k | IBM | 1* | -- | 115 | 0.80 | 5.16 | |
Sarvam 30B (high) | 65.5k | Sarvam | 1* | -- | -- | -- | -- | |
Gemma 3 270M | 32k | Google | 1* | -- | -- | -- | -- | |
Qwen3.5 2B (Non-reasoning) | 262k | Alibaba | 1* | -- | -- | -- | -- | |
Qwen3 Omni 30B A3B | 65.5k | Alibaba | 1* | -- | 92 | 1.89 | 7.35 | |
Granite 4.0 Micro | 128k | IBM | 1* | -- | -- | -- | -- | |
Phi-4 Multimodal | 128k |  Microsoft | 1* | -- | 17 | 0.86 | 30.48 | |
EXAONE 4.0 32B (Non-reasoning) | 131k | LG AI Research | 1* | -- | -- | -- | -- | |
Olmo 3 7B Think | 65.5k | Allen Institute for AI | 1* | -- | -- | -- | -- | |
Nova Micro | 130k | Amazon | 1* | -- | 278 | 0.87 | 2.67 | |
LFM2.5-VL-1.6B | 32k | Liquid AI | 1* | -- | 358 | 1.59 | 2.98 | |
Molmo2-8B | 36.9k | Allen Institute for AI | 1* | -- | -- | -- | -- | |
Qwen3.5 0.8B (Non-reasoning) | 262k | Alibaba | 1* | -- | -- | -- | -- | |
MiniCPM-V 4.6 1.3B | 262k | OpenBMB | 1* | -- | -- | -- | -- | |
LFM2.5-1.2B-Instruct | 32k | Liquid AI | 1* | -- | -- | -- | -- | |
Apertus 70B Instruct | 65.5k |  Swiss AI Initiative | 1* | -- | -- | -- | -- | |
Granite 4.0 350M | 32.8k | IBM | 1* | -- | -- | -- | -- | |
LFM2.5-1.2B-Thinking | 32k | Liquid AI | 1* | -- | -- | -- | -- | |
Granite 4.0 H Small | 128k | IBM | 1* | -- | 26 | 11.42 | 30.40 | |
Llama 3.2 11B (Vision) | 128k | Meta | 1* | -- | 27 | 1.10 | 19.78 | |
LFM2 8B A1B | 32.8k | Liquid AI | 1* | -- | -- | -- | -- | |
LFM2 2.6B | 32.8k | Liquid AI | 1* | -- | -- | -- | -- | |
Gemma 4 E2B (Non-reasoning) | 128k | Google | 1* | -- | -- | -- | -- | |
Jamba 1.7 Large | 256k |  AI21 Labs | 1* | -- | -- | -- | -- | |
Tiny Aya Global | 8.19k | Cohere | 1* | -- | -- | -- | -- | |
Exaone 4.0 1.2B (Non-reasoning) | 64k | LG AI Research | 1* | -- | -- | -- | -- | |
Apertus 8B Instruct | 65.5k | Swiss AI Initiative | 1* | -- | -- | -- | -- | |
Olmo 3 7B | 65.5k | Allen Institute for AI | 1* | -- | -- | -- | -- | |
Granite 4.1 3B | 131k | IBM | 1* | -- | -- | -- | -- | |
Olmo 3.1 32B Instruct | 65.5k | Allen Institute for AI | 1* | -- | -- | -- | -- | |
R1 1776 | 128k |  Perplexity | 1* | -- | -- | -- | -- | |
Llama 3.2 90B (Vision) | 128k | Meta | 1* | -- | -- | -- | -- | |
Phi-4 Mini | 128k | Microsoft | 1* | -- | 44 | 0.86 | 12.35 | |
Granite 4.0 1B | 128k | IBM | 1* | -- | -- | -- | -- | |
Granite 4.0 H 350M | 32.8k | IBM | 1* | -- | -- | -- | -- | |
Qwen3.5 0.8B | 262k | Alibaba | 1* | -- | -- | -- | -- | |
DeepHermes 3 - Mistral 24B | 32k | Nous Research | 1* | -- | -- | -- | -- | |
NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning) | 128k | NVIDIA | 1* | -- | 119 | 2.29 | 6.48 | |
Molmo 7B-D | 4.1k | Allen Institute for AI | 1* | -- | -- | -- | -- | |
Jamba 1.7 Mini | 258k | AI21 Labs | 1* | -- | -- | -- | -- | |
DeepHermes 3 - Llama-3.1 8B | 128k | Nous Research | 1* | -- | -- | -- | -- | |
Exaone 4.0 1.2B | 64k | LG AI Research | 1* | -- | -- | -- | -- | |
Granite 4.0 H 1B | 128k | IBM | 1* | -- | -- | -- | -- | |
Reka Flash 3 | 128k |  Reka AI | 1* | -- | -- | -- | -- | |
Phi-4 | 16k | Microsoft | 1* | -- | 37 | 2.57 | 16.26 | |
Jamba Reasoning 3B | 262k | AI21 Labs | 1* | -- | -- | -- | -- | |
LFM2 24B A2B | 32.8k | Liquid AI | 1* | -- | -- | -- | -- | |
Claude Sonnet 5 (low) | 1M | Anthropic | -- | -- | 61 | 1.83 | 10.04 | |
EXAONE 4.5 33B (Non-reasoning) | 262k | LG AI Research | -- | -- | -- | -- | -- | |
Claude Sonnet 5 (xhigh) | 1M | Anthropic | -- | -- | 84 | 31.80 | 37.77 | |
Gemini 3 Deep Think | 128k | Google | -- | -- | -- | -- | -- | |
Claude Sonnet 5 (high) | 1M | Anthropic | -- | -- | 69 | 10.98 | 18.19 | |
GPT-5.5 Pro (xhigh) | 922k | OpenAI | -- | -- | -- | -- | -- | |
Cogito v2.1 | 128k |  Deep Cogito | -- | -- | -- | -- | -- | |
Claude Sonnet 5 (medium) | 1M | Anthropic | -- | -- | 64 | 3.08 | 10.85 | |
Definiciones clave
Preguntas frecuentes
Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) ocupa actualmente el puesto #1 en el ranking de modelos LLM de Artificial Analysis, con una puntuación de 57 en el Índice de Inteligencia, entre 54 modelos clasificados.
Los mejores modelos por Índice de Inteligencia son: 1. Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) (57), 2. GPT-6 Astra (max) (55), 3. GPT-6 Astra (xhigh) (54), 4. Claude Opus 5 (Adaptive Reasoning, Max Effort) (54), 5. Claude Opus 5 (Adaptive Reasoning, Xhigh Effort) (53).
Celeris-1 es el más rápido con 1,530.4 tokens por segundo, seguido por Mercury 2 (820.6 t/s) y Gemini 2.5 Flash-Lite (Reasoning) (372.7 t/s).
Granite 4.2 3B tiene el menor costo por tarea del Índice de Inteligencia con $0.01, seguido por Granite 4.2 8B ($0.02) y Ling 3.0 Flash ($0.03).
Kimi K3 (max) es el modelo con pesos abiertos mejor clasificado, con una puntuación de 50 en el Índice de Inteligencia. Hay 18 modelos con pesos abiertos de un total de 54 en el ranking.
Los mejores modelos con pesos abiertos por Índice de Inteligencia son: 1. Kimi K3 (max) (50), 2. GLM-5.3 (max) (49), 3. Qwen3.8 2.4T A95B (47).
Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) lidera entre 52 modelos de razonamiento, con una puntuación de 57 en el Índice de Inteligencia. Los modelos de razonamiento usan pensamiento extendido para resolver problemas complejos antes de responder.
El ranking incluye filtros para limitar los resultados por tipo de modelo (razonamiento o no razonamiento), apertura (pesos abiertos o propietario) y otros criterios. También puedes ajustar las opciones de prompt para ver cómo varía el rendimiento con distintas longitudes de entrada.
Haz clic en cualquier nombre de modelo en la clasificación para visitar su página de comparación dedicada, con gráficos detallados sobre inteligencia, precios, velocidad, latencia y más. También puedes comparar proveedores de API para cada modelo. Ver todos los modelos