o4-mini (high) |  | Proprietary | 200k | 70 | 0.83% | 0.78% | 0.18% | 0.8% | 0.47% | 0.99% | 0.99% | 0.94% | 0% | 0% | $1.93/1M | 4.4/1M | 1.1/1M | 126.5t/s | 77.1t/s | 105.3t/s | 148t/s | 187.4t/s | 39.16t/s | 18.64t/s | 28.35t/s | 56.43t/s | 79.18t/s | |
Gemini 2.5 Pro Preview |  | Proprietary | 1m | 68 | 0.86% | 0.84% | 0.17% | 0.7% | 0.4% | 0.99% | 0.98% | 0.87% | 0% | 0% | $3.44/1M | 10.0/1M | 1.25/1M | 220.5t/s | 196.1t/s | 212.7t/s | 228.4t/s | 234.1t/s | 27.4t/s | 12.91t/s | 17.47t/s | 30.22t/s | 34.85t/s | |
Grok 3 mini Reasoning (high) |  | Proprietary | 1m | 67 | 0.83% | 0.79% | 0.11% | 0.7% | 0.41% | 0.98% | 0.99% | 0.93% | 0% | 0% | $0.35/1M | 0.5/1M | 0.3/1M | 112.5t/s | 86.6t/s | 98.7t/s | 141.5t/s | 212.8t/s | 0.36t/s | 0.3t/s | 0.33t/s | 0.43t/s | 0.5t/s | |
o3-mini (high) |  | Proprietary | 200k | 66 | 0.8% | 0.77% | 0.12% | 0.73% | 0.4% | 0% | 0.99% | 0.86% | 0% | 0% | $1.93/1M | 4.4/1M | 1.1/1M | 208.2t/s | 99.9t/s | 185.4t/s | 225.3t/s | 278.5t/s | 31.67t/s | 19.82t/s | 27.54t/s | 36.86t/s | 48.16t/s | |
o3-mini |  | Proprietary | 200k | 63 | 0.79% | 0.75% | 0.09% | 0.72% | 0.4% | 0.97% | 0.97% | 0.77% | 0% | 0% | $1.93/1M | 4.4/1M | 1.1/1M | 219t/s | 151.2t/s | 197.2t/s | 249.9t/s | 302.1t/s | 10.7t/s | 6.75t/s | 9.25t/s | 13.08t/s | 18.01t/s | |
o1 |  | Proprietary | 200k | 62 | 0.84% | 0.75% | 0.08% | 0.68% | 0.36% | 0.97% | 0.97% | 0.72% | 0.88% | 0% | $26.25/1M | 60.0/1M | 15.0/1M | 86.6t/s | 57t/s | 70.9t/s | 97.5t/s | 118.7t/s | 33.95t/s | 21.22t/s | 30.97t/s | 38.1t/s | 60.95t/s | |
DeepSeek R1 |  | Open | 128k | 60 | 0.84% | 0.71% | 0.09% | 0.62% | 0.36% | 0.98% | 0.97% | 0.68% | 0% | 0% | $0.96/1M | 2.19/1M | 0.55/1M | 25.3t/s | 21.4t/s | 23.4t/s | 30t/s | 35.1t/s | 3.38t/s | 2.82t/s | 3.11t/s | 4.5t/s | 6.36t/s | |
QwQ-32B |  | Open | 131k | 58 | 0.76% | 0.59% | 0.08% | 0.63% | 0.36% | 0.98% | 0.96% | 0.78% | 0% | 0% | $0.47/1M | 0.55/1M | 0.4/1M | 89.1t/s | 21t/s | 40.9t/s | 361.2t/s | 429.7t/s | 0.49t/s | 0.19t/s | 0.29t/s | 1.33t/s | 3.31t/s | |
Claude 3.7 Sonnet Thinking |  | Proprietary | 200k | 57 | 0.84% | 0.77% | 0.1% | 0.47% | 0.4% | 0.98% | 0.95% | 0.49% | 0% | 0% | $6.0/1M | 15.0/1M | 3.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
o1-mini |  | Proprietary | 128k | 54 | 0.74% | 0.6% | 0.05% | 0.58% | 0.32% | 0.97% | 0.94% | 0.6% | 0.83% | 1308% | $1.93/1M | 4.4/1M | 1.1/1M | 215.4t/s | 104t/s | 179.7t/s | 222.8t/s | 235.4t/s | 9.68t/s | 1.4t/s | 5.55t/s | 11.5t/s | 14.12t/s | |
GPT-4.1 mini |  | Proprietary | 1m | 53 | 0.78% | 0.66% | 0.05% | 0.48% | 0.4% | 0.95% | 0.93% | 0.43% | 0% | 0% | $0.7/1M | 1.6/1M | 0.4/1M | 84.2t/s | 67.4t/s | 75.5t/s | 93.2t/s | 117.6t/s | 0.35t/s | 0.26t/s | 0.32t/s | 0.41t/s | 0.58t/s | |
DeepSeek V3 (Mar' 25) |  | Open | 128k | 53 | 0.82% | 0.66% | 0.05% | 0.41% | 0.36% | 0.92% | 0.94% | 0.52% | 0% | 0% | $0.48/1M | 1.1/1M | 0.27/1M | 26t/s | 23.6t/s | 24.3t/s | 28.9t/s | 34.1t/s | 3.68t/s | 3.08t/s | 3.27t/s | 4t/s | 4.26t/s | |
GPT-4.1 |  | Proprietary | 1m | 53 | 0.81% | 0.67% | 0.05% | 0.46% | 0.38% | 0.96% | 0.91% | 0.44% | 0% | 0% | $3.5/1M | 8.0/1M | 2.0/1M | 130.8t/s | 66.8t/s | 102.3t/s | 144.7t/s | 175.3t/s | 0.4t/s | 0.34t/s | 0.36t/s | 0.56t/s | 0.8t/s | |
DeepSeek R1 Distill Qwen 32B |  | Open | 128k | 52 | 0.74% | 0.62% | 0.06% | 0.27% | 0.38% | 0.95% | 0.94% | 0.69% | 0% | 0% | $0.22/1M | 0.24/1M | 0.21/1M | 33.8t/s | 18.3t/s | 20.7t/s | 46.3t/s | 49.9t/s | 1.02t/s | 0.2t/s | 0.24t/s | 1.12t/s | 1.23t/s | |
Gemini 2.0 Flash Thinking exp. (Jan '25) |  | Proprietary | 1m | 52 | 0.8% | 0.7% | 0.07% | 0.32% | 0.33% | 0% | 0.94% | 0.5% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Llama 4 Maverick |  | Open | 1m | 51 | 0.81% | 0.67% | 0.05% | 0.4% | 0.33% | 0.88% | 0.89% | 0.39% | 0% | 0% | $0.36/1M | 0.82/1M | 0.2/1M | 123.9t/s | 54.2t/s | 94t/s | 136t/s | 806.5t/s | 0.4t/s | 0.16t/s | 0.34t/s | 0.53t/s | 0.92t/s | |
Grok 3 |  | Proprietary | 1m | 51 | 0.8% | 0.69% | 0.05% | 0.43% | 0.37% | 0.91% | 0.87% | 0.33% | 0% | 0% | $6.0/1M | 15.0/1M | 3.0/1M | 53.9t/s | 32.6t/s | 44.6t/s | 59t/s | 66.7t/s | 0.55t/s | 0.47t/s | 0.52t/s | 0.68t/s | 3.8t/s | |
GPT-4o (March 2025) |  | Proprietary | 128k | 50 | 0.8% | 0.66% | 0.05% | 0.43% | 0.37% | 0.96% | 0.89% | 0.33% | 0% | 0% | $7.5/1M | 15.0/1M | 5.0/1M | 211.3t/s | 120.5t/s | 194t/s | 229.9t/s | 240.7t/s | 0.56t/s | 0.35t/s | 0.41t/s | 0.63t/s | 1t/s | |
DeepSeek R1 Distill Qwen 14B |  | Open | 128k | 49 | 0.74% | 0.48% | 0.04% | 0.38% | 0.24% | 0.93% | 0.95% | 0.67% | 0% | 0% | $0.88/1M | 0.88/1M | 0.88/1M | 72.7t/s | 41.8t/s | 44.8t/s | 152.9t/s | 171t/s | 0.61t/s | 0.27t/s | 0.38t/s | 0.74t/s | 0.86t/s | |
Gemini 2.0 Pro Experimental |  | Proprietary | 2m | 49 | 0.81% | 0.62% | 0.07% | 0.35% | 0.31% | 0.95% | 0.92% | 0.36% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 218.3t/s | 203.3t/s | 211.6t/s | 225.1t/s | 233.8t/s | 27.15t/s | 12.32t/s | 17.21t/s | 30.74t/s | 36.09t/s | |
Gemini 2.0 Flash |  | Proprietary | 1m | 48 | 0.78% | 0.62% | 0.05% | 0.33% | 0.31% | 0.9% | 0.93% | 0.33% | 0% | 0% | $0.17/1M | 0.4/1M | 0.1/1M | 255.2t/s | 230.6t/s | 243.5t/s | 265.3t/s | 284.3t/s | 0.33t/s | 0.25t/s | 0.28t/s | 0.4t/s | 0.41t/s | |
Claude 3.7 Sonnet |  | Proprietary | 200k | 48 | 0.8% | 0.66% | 0.05% | 0.39% | 0.38% | 0.95% | 0.85% | 0.22% | 0% | 0% | $6.0/1M | 15.0/1M | 3.0/1M | 78.4t/s | 65.4t/s | 74.2t/s | 79.7t/s | 83.1t/s | 1.01t/s | 0.71t/s | 0.81t/s | 1.53t/s | 2.52t/s | |
DeepSeek R1 Distill Llama 70B |  | Open | 128k | 48 | 0.8% | 0.4% | 0.06% | 0.27% | 0.31% | 0.97% | 0.94% | 0.67% | 0% | 0% | $0.6/1M | 0.87/1M | 0.54/1M | 83.3t/s | 29.3t/s | 56.6t/s | 347.4t/s | 2313.00t/s | 0.5t/s | 0.17t/s | 0.34t/s | 0.76t/s | 1.72t/s | |
Reka Flash 3 |  | Open | 128k | 47 | 0.67% | 0.53% | 0.05% | 0.44% | 0.27% | 0.95% | 0.89% | 0.51% | 0% | 0% | $0.35/1M | 0.8/1M | 0.2/1M | 56.2t/s | 55.2t/s | 55.5t/s | 57.1t/s | 58t/s | 0.96t/s | 0.87t/s | 0.95t/s | 0.98t/s | 1.07t/s | |
DeepSeek V3 (Dec '24) |  | Open | 128k | 46 | 0.75% | 0.56% | 0.04% | 0.36% | 0.35% | 0.91% | 0.89% | 0.25% | 0.86% | 0% | $0.48/1M | 1.1/1M | 0.27/1M | 26.2t/s | 23.7t/s | 24.4t/s | 29.6t/s | 35.7t/s | 3.66t/s | 3.05t/s | 3.21t/s | 3.9t/s | 4.78t/s | |
Gemini 2.0 Flash (exp) |  | Proprietary | 1m | 46 | 0.78% | 0.64% | 0.05% | 0.21% | 0.34% | 0.91% | 0.91% | 0.3% | 0.84% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 250.8t/s | 227.9t/s | 241.9t/s | 257.3t/s | 268.1t/s | 0.27t/s | 0.23t/s | 0.25t/s | 0.29t/s | 0.36t/s | |
Gemini 1.5 Pro (Sep) |  | Proprietary | 2m | 45 | 0.75% | 0.59% | 0.05% | 0.32% | 0.3% | 0.9% | 0.88% | 0.23% | 0.85% | 1301% | $2.19/1M | 5.0/1M | 1.25/1M | 99.6t/s | 90.3t/s | 94.7t/s | 103.9t/s | 106.6t/s | 0.46t/s | 0.41t/s | 0.44t/s | 0.54t/s | 0.92t/s | |
Qwen2.5 Max |  | Proprietary | 32k | 45 | 0.76% | 0.59% | 0.05% | 0.36% | 0.34% | 0.93% | 0.84% | 0.23% | 0% | 0% | $2.8/1M | 6.4/1M | 1.6/1M | 52.4t/s | 45.9t/s | 50.7t/s | 53.6t/s | 54t/s | 1.27t/s | 1.14t/s | 1.19t/s | 1.37t/s | 1.49t/s | |
Claude 3.5 Sonnet (Oct) |  | Proprietary | 200k | 44 | 0.77% | 0.6% | 0.04% | 0.38% | 0.37% | 0.93% | 0.77% | 0.16% | 0.88% | 1282% | $6.0/1M | 15.0/1M | 3.0/1M | 80.3t/s | 68.1t/s | 77.7t/s | 81.5t/s | 83.3t/s | 0.87t/s | 0.65t/s | 0.78t/s | 1.22t/s | 1.64t/s | |
QwQ 32B-Preview |  | Open | 33k | 43 | 0.65% | 0.56% | 0.05% | 0.34% | 0.04% | 0.87% | 0.91% | 0.45% | 0% | 0% | $0.58/1M | 0.75/1M | 0.55/1M | 54.6t/s | 38.4t/s | 46.2t/s | 82.5t/s | 98t/s | 0.41t/s | 0.2t/s | 0.25t/s | 0.56t/s | 1.68t/s | |
Sonar Pro |  | Proprietary | 200k | 43 | 0.76% | 0.58% | 0.08% | 0.28% | 0.23% | 0.85% | 0.75% | 0.29% | 0% | 0% | $6.0/1M | 15.0/1M | 3.0/1M | 105.5t/s | 50t/s | 77.8t/s | 137.8t/s | 178.5t/s | 2.45t/s | 1.71t/s | 1.96t/s | 2.87t/s | 3.84t/s | |
Llama 4 Scout |  | Open | 10m | 43 | 0.75% | 0.59% | 0.04% | 0.3% | 0.17% | 0.83% | 0.84% | 0.28% | 0% | 0% | $0.26/1M | 0.5/1M | 0.15/1M | 121.3t/s | 39.2t/s | 89.5t/s | 155.8t/s | 2733.30t/s | 0.41t/s | 0.2t/s | 0.32t/s | 0.52t/s | 0.83t/s | |
Sonar |  | Proprietary | 127k | 43 | 0.69% | 0.47% | 0.07% | 0.3% | 0.23% | 0.82% | 0.82% | 0.49% | 0% | 0% | $1.0/1M | 1.0/1M | 1.0/1M | 77.1t/s | 35.5t/s | 57.7t/s | 86.6t/s | 118.9t/s | 2.28t/s | 1.78t/s | 2.02t/s | 2.69t/s | 3.51t/s | |
Llama 3.3 70B |  | Open | 128k | 41 | 0.71% | 0.5% | 0.04% | 0.29% | 0.26% | 0.86% | 0.77% | 0.3% | 0.84% | 0% | $0.6/1M | 0.7/1M | 0.59/1M | 121t/s | 25.7t/s | 45.9t/s | 184t/s | 2104.30t/s | 0.47t/s | 0.24t/s | 0.35t/s | 0.6t/s | 1.49t/s | |
Gemini 2.0 Flash-Lite (Feb '25) |  | Proprietary | 1m | 41 | 0.72% | 0.54% | 0.04% | 0.19% | 0.25% | 0.88% | 0.87% | 0.28% | 0% | 0% | $0.13/1M | 0.3/1M | 0.07/1M | 203.7t/s | 178.7t/s | 193.5t/s | 210.5t/s | 216t/s | 0.29t/s | 0.25t/s | 0.28t/s | 0.31t/s | 0.35t/s | |
GPT-4o (Nov '24) |  | Proprietary | 128k | 41 | 0.75% | 0.54% | 0.03% | 0.31% | 0.33% | 0.93% | 0.76% | 0.15% | 0.84% | 1361% | $4.38/1M | 10.0/1M | 2.5/1M | 174.6t/s | 107.8t/s | 144.1t/s | 200.7t/s | 219t/s | 0.61t/s | 0.46t/s | 0.51t/s | 0.66t/s | 0.8t/s | |
GPT-4o (May '24) |  | Proprietary | 128k | 41 | 0.74% | 0.53% | 0.03% | 0.33% | 0.31% | 0.94% | 0.79% | 0.11% | 0% | 1285% | $7.5/1M | 15.0/1M | 5.0/1M | 124.4t/s | 77.5t/s | 89.6t/s | 138.7t/s | 154.7t/s | 0.52t/s | 0.42t/s | 0.47t/s | 0.61t/s | 0.85t/s | |
GPT-4.1 nano |  | Proprietary | 1m | 41 | 0.66% | 0.51% | 0.04% | 0.33% | 0.26% | 0.88% | 0.85% | 0.24% | 0% | 0% | $0.17/1M | 0.4/1M | 0.1/1M | 242.1t/s | 158.6t/s | 218.6t/s | 294.6t/s | 360.9t/s | 0.46t/s | 0.25t/s | 0.34t/s | 0.86t/s | 0.9t/s | |
Tulu3 405B |  | Open | 128k | 40 | 0.72% | 0.52% | 0.04% | 0.29% | 0.3% | 0.89% | 0.78% | 0.13% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Command A |  | Open | 256k | 40 | 0.71% | 0.53% | 0.05% | 0.29% | 0.28% | 0.82% | 0.82% | 0.1% | 0% | 0% | $4.38/1M | 10.0/1M | 2.5/1M | 55.7t/s | 49.4t/s | 53.2t/s | 57t/s | 58.8t/s | 0.26t/s | 0.23t/s | 0.24t/s | 0.26t/s | 0.38t/s | |
Phi-4 |  | Open | 16k | 40 | 0.71% | 0.57% | 0.04% | 0.23% | 0.26% | 0.87% | 0.81% | 0.14% | 0% | 0% | $0.22/1M | 0.5/1M | 0.13/1M | 45.2t/s | 39.1t/s | 44t/s | 45.5t/s | 45.5t/s | 0.48t/s | 0.4t/s | 0.45t/s | 0.52t/s | 18.07t/s | |
MiniMax-Text-01 |  | Open | 4m | 40 | 0.76% | 0.58% | 0.04% | 0.25% | 0.25% | 0.86% | 0.75% | 0.13% | 0% | 0% | $0.42/1M | 1.1/1M | 0.2/1M | 28.8t/s | 20.6t/s | 23.8t/s | 33.7t/s | 39.7t/s | 0.85t/s | 0.63t/s | 0.72t/s | 1.03t/s | 1.53t/s | |
Qwen2.5 72B |  | Open | 131k | 40 | 0.72% | 0.49% | 0.04% | 0.28% | 0.27% | 0.88% | 0.86% | 0.16% | 0.83% | 1259% | $0.0/1M | 0.0/1M | 0.0/1M | 63.1t/s | 59t/s | 62.7t/s | 63.2t/s | 63.5t/s | 1.17t/s | 1.02t/s | 1.12t/s | 1.27t/s | 1.49t/s | |
Llama 3.1 405B |  | Open | 128k | 40 | 0.73% | 0.52% | 0.04% | 0.31% | 0.3% | 0.85% | 0.7% | 0.21% | 0.77% | 1266% | $3.5/1M | 3.5/1M | 3.5/1M | 31.8t/s | 11.8t/s | 27t/s | 78.3t/s | 171.4t/s | 0.78t/s | 0.37t/s | 0.46t/s | 1.35t/s | 2.15t/s | |
Llama 3.3 Nemotron Super 49B v1 |  | Open | 128k | 39 | 0.7% | 0.52% | 0.04% | 0.28% | 0.23% | 0.83% | 0.78% | 0.19% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Gemini 1.5 Flash (Sep) |  | Proprietary | 1m | 39 | 0.68% | 0.46% | 0.04% | 0.27% | 0.27% | 0.84% | 0.83% | 0.18% | 0.81% | 1271% | $0.13/1M | 0.3/1M | 0.07/1M | 198t/s | 187.3t/s | 194.1t/s | 201.2t/s | 207.2t/s | 0.21t/s | 0.19t/s | 0.2t/s | 0.22t/s | 0.32t/s | |
Grok 2 |  | Proprietary | 131k | 39 | 0.71% | 0.51% | 0.04% | 0.27% | 0.28% | 0.86% | 0.78% | 0.13% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Grok Beta |  | Proprietary | 128k | 38 | 0.7% | 0.47% | 0.05% | 0.24% | 0.3% | 0.87% | 0.74% | 0.1% | 0% | 1289% | $7.5/1M | 15.0/1M | 5.0/1M | 65.4t/s | 56.4t/s | 61.8t/s | 66.4t/s | 68.2t/s | 0.28t/s | 0.24t/s | 0.26t/s | 0.32t/s | 0.39t/s | |
Gemma 3 27B |  | Open | 128k | 38 | 0.67% | 0.43% | 0.05% | 0.14% | 0.21% | 0.89% | 0.88% | 0.25% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Mistral Large 2 (Nov '24) |  | Open | 128k | 38 | 0.7% | 0.49% | 0.04% | 0.29% | 0.29% | 0.9% | 0.74% | 0.11% | 0.83% | 0% | $3.0/1M | 6.0/1M | 2.0/1M | 36.6t/s | 15.3t/s | 29.9t/s | 41.2t/s | 46.3t/s | 0.39t/s | 0.36t/s | 0.37t/s | 0.44t/s | 1.32t/s | |
Llama 3.1 Nemotron 70B |  | Open | 128k | 37 | 0.69% | 0.47% | 0.05% | 0.17% | 0.23% | 0.82% | 0.73% | 0.25% | 0% | 1269% | $0.24/1M | 0.4/1M | 0.18/1M | 41t/s | 32.5t/s | 39.2t/s | 64.8t/s | 71.7t/s | 0.53t/s | 0.26t/s | 0.44t/s | 0.63t/s | 0.87t/s | |
Qwen2.5 Instruct 32B |  | Open | 128k | 37 | 0.7% | 0.47% | 0.04% | 0.25% | 0.23% | 0.9% | 0.81% | 0.11% | 0% | 0% | $0.15/1M | 0.3/1M | 0.1/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Pixtral Large |  | Open | 128k | 37 | 0.7% | 0.51% | 0.04% | 0.26% | 0.29% | 0.85% | 0.71% | 0.07% | 0% | 0% | $3.0/1M | 6.0/1M | 2.0/1M | 28.3t/s | 14.4t/s | 21.8t/s | 37.4t/s | 41.4t/s | 0.48t/s | 0.36t/s | 0.4t/s | 0.59t/s | 9.39t/s | |
Mistral Large 2 (Jul '24) |  | Open | 128k | 37 | 0.68% | 0.47% | 0.03% | 0.27% | 0.27% | 0.89% | 0.71% | 0.09% | 0% | 1251% | $3.0/1M | 6.0/1M | 2.0/1M | 35.8t/s | 10.2t/s | 28.3t/s | 40.6t/s | 43.4t/s | 0.46t/s | 0.39t/s | 0.41t/s | 0.71t/s | 2.85t/s | |
Nova Pro |  | Proprietary | 300k | 37 | 0.69% | 0.5% | 0.03% | 0.23% | 0.21% | 0.83% | 0.79% | 0.11% | 0.83% | 0% | $1.75/1M | 4.0/1M | 1.0/1M | 113.2t/s | 98.4t/s | 109.2t/s | 116.4t/s | 123.5t/s | 0.4t/s | 0.36t/s | 0.38t/s | 0.43t/s | 0.53t/s | |
GPT-4o mini |  | Proprietary | 128k | 36 | 0.65% | 0.43% | 0.04% | 0.23% | 0.23% | 0.88% | 0.79% | 0.12% | 0.80% | 1273% | $0.26/1M | 0.6/1M | 0.15/1M | 88.4t/s | 60.6t/s | 74t/s | 99.9t/s | 116.1t/s | 0.34t/s | 0.28t/s | 0.31t/s | 0.38t/s | 0.55t/s | |
Qwen2.5 Coder 32B |  | Open | 131k | 36 | 0.64% | 0.42% | 0.04% | 0.3% | 0.27% | 0.9% | 0.77% | 0.12% | 0% | 1220% | $0.2/1M | 0.2/1M | 0.2/1M | 52.9t/s | 13.8t/s | 41.2t/s | 67.8t/s | 87.6t/s | 0.46t/s | 0.21t/s | 0.29t/s | 0.66t/s | 1.64t/s | |
Claude 3 Opus |  | Proprietary | 200k | 35 | 0.7% | 0.49% | 0.03% | 0.28% | 0.23% | 0.85% | 0.64% | 0.03% | 0% | 1248% | $30.0/1M | 75.0/1M | 15.0/1M | 29.1t/s | 25.3t/s | 27.2t/s | 30.1t/s | 31.5t/s | 1.06t/s | 0.88t/s | 0.99t/s | 1.14t/s | 1.75t/s | |
Mistral Small 3 |  | Open | 32k | 35 | 0.65% | 0.46% | 0.04% | 0.25% | 0.24% | 0.85% | 0.72% | 0.08% | 0% | 0% | $0.15/1M | 0.3/1M | 0.1/1M | 139.3t/s | 69.8t/s | 125t/s | 156t/s | 165.5t/s | 0.27t/s | 0.25t/s | 0.27t/s | 0.3t/s | 1.2t/s | |
Mistral Small 3.1 |  | Open | 128k | 35 | 0.66% | 0.45% | 0.05% | 0.21% | 0.27% | 0.86% | 0.71% | 0.09% | 0% | 0% | $0.15/1M | 0.3/1M | 0.1/1M | 147.9t/s | 68.2t/s | 106.1t/s | 162t/s | 179.5t/s | 0.28t/s | 0.24t/s | 0.26t/s | 0.31t/s | 0.41t/s | |
Llama 3.1 70B |  | Open | 128k | 35 | 0.68% | 0.41% | 0.05% | 0.23% | 0.27% | 0.81% | 0.65% | 0.17% | 0% | 1249% | $0.66/1M | 0.73/1M | 0.66/1M | 47.7t/s | 10.7t/s | 31.7t/s | 108.1t/s | 161.7t/s | 0.54t/s | 0.23t/s | 0.41t/s | 0.78t/s | 7.22t/s | |
Claude 3.5 Haiku |  | Proprietary | 200k | 35 | 0.63% | 0.41% | 0.04% | 0.31% | 0.27% | 0.86% | 0.72% | 0.03% | 78% | 0% | $1.6/1M | 4.0/1M | 0.8/1M | 66.1t/s | 60.7t/s | 64.4t/s | 68t/s | 69.2t/s | 5.91t/s | 1.59t/s | 3.83t/s | 7.37t/s | 9.13t/s | |
Qwen Turbo |  | Proprietary | 1m | 34 | 0.63% | 0.41% | 0.04% | 0.16% | 0.15% | 0.85% | 0.81% | 0.12% | 0% | 0% | $0.09/1M | 0.2/1M | 0.05/1M | 110.4t/s | 96.4t/s | 106.9t/s | 111.6t/s | 114.9t/s | 1.11t/s | 0.96t/s | 0.99t/s | 1.19t/s | 1.38t/s | |
Gemini 1.5 Pro (May) |  | Proprietary | 2m | 34 | 0.66% | 0.37% | 0.04% | 0.24% | 0.27% | 0.83% | 0.67% | 0.08% | 0% | 1260% | $2.19/1M | 5.0/1M | 1.25/1M | 68.1t/s | 63.7t/s | 66.1t/s | 69.4t/s | 70.7t/s | 0.37t/s | 0.34t/s | 0.35t/s | 0.39t/s | 0.46t/s | |
Gemma 3 12B |  | Open | 128k | 34 | 0.6% | 0.35% | 0.05% | 0.14% | 0.17% | 0.83% | 0.85% | 0.22% | 0% | 0% | $0.06/1M | 0.1/1M | 0.05/1M | 32.4t/s | 19.9t/s | 26.3t/s | 42.9t/s | 53.5t/s | 0.53t/s | 0.4t/s | 0.45t/s | 0.67t/s | 0.89t/s | |
DeepSeek R1 Distill Llama 8B |  | Open | 128k | 34 | 0.54% | 0.3% | 0.04% | 0.23% | 0.12% | 0.84% | 0.85% | 0.33% | 0% | 0% | $0.04/1M | 0.04/1M | 0.04/1M | 48.2t/s | 25.9t/s | 34t/s | 53.6t/s | 56.1t/s | 0.81t/s | 0.6t/s | 0.68t/s | 15.94t/s | 50.16t/s | |
Nova Lite |  | Proprietary | 300k | 33 | 0.59% | 0.43% | 0.05% | 0.17% | 0.14% | 0.84% | 0.77% | 0.11% | 0.76% | 0% | $0.1/1M | 0.24/1M | 0.06/1M | 282.7t/s | 244.3t/s | 264.1t/s | 305.4t/s | 323.9t/s | 0.33t/s | 0.3t/s | 0.33t/s | 0.34t/s | 0.38t/s | |
Qwen2 72B |  | Open | 131k | 33 | 0.62% | 0.37% | 0.04% | 0.16% | 0.23% | 0.83% | 0.7% | 0.15% | 0% | 1187% | $0.0/1M | 0.0/1M | 0.0/1M | 31t/s | 30.8t/s | 30.9t/s | 31.1t/s | 31.2t/s | 1.36t/s | 1.22t/s | 1.29t/s | 1.44t/s | 1.75t/s | |
Llama 3.2 90B (Vision) |  | Open | 128k | 33 | 0.67% | 0.43% | 0.05% | 0.21% | 0.24% | 0.82% | 0.63% | 0.05% | 0% | 0% | $0.81/1M | 0.81/1M | 0.81/1M | 39.9t/s | 24.6t/s | 32t/s | 45.8t/s | 60.9t/s | 0.35t/s | 0.17t/s | 0.21t/s | 0.48t/s | 0.61t/s | |
Gemini 1.5 Flash-8B |  | Proprietary | 1m | 31 | 0.57% | 0.36% | 0.05% | 0.22% | 0.23% | 0.12% | 0.69% | 0.03% | 0.74% | 1211% | $0.07/1M | 0.15/1M | 0.04/1M | 285.9t/s | 273.6t/s | 280t/s | 293.8t/s | 302.7t/s | 0.19t/s | 0.16t/s | 0.18t/s | 0.23t/s | 0.26t/s | |
DeepHermes 3 - Mistral 24B |  | Open | 32k | 30 | 0.58% | 0.38% | 0.04% | 0.2% | 0.23% | 0.75% | 0.6% | 0.05% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Jamba 1.6 Large |  | Open | 256k | 29 | 0.56% | 0.39% | 0.04% | 0.17% | 0.18% | 0.7% | 0.58% | 0.05% | 0% | 0% | $3.5/1M | 8.0/1M | 2.0/1M | 66.2t/s | 42.7t/s | 62.4t/s | 68.8t/s | 70.6t/s | 0.57t/s | 0.37t/s | 0.49t/s | 0.68t/s | 1.02t/s | |
Hermes 3 - Llama-3.1 70B |  | Open | 128k | 29 | 0.57% | 0.4% | 0.04% | 0.19% | 0.23% | 0.75% | 0.54% | 0.02% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Jamba 1.5 Large |  | Open | 256k | 29 | 0.57% | 0.43% | 0.04% | 0.14% | 0.16% | 0.24% | 0.61% | 0.05% | 0% | 1221% | $3.5/1M | 8.0/1M | 2.0/1M | 68t/s | 52.3t/s | 66.8t/s | 68.3t/s | 69.5t/s | 0.5t/s | 0.35t/s | 0.39t/s | 0.58t/s | 0.91t/s | |
Yi-Large |  | Proprietary | 32k | 28 | 0.59% | 0.36% | 0.03% | 0.11% | 0.19% | 0.74% | 0.56% | 0.07% | 0% | 1213% | $3.0/1M | 3.0/1M | 3.0/1M | 69t/s | 63.7t/s | 66.5t/s | 79.2t/s | 81.8t/s | 0.39t/s | 0.32t/s | 0.36t/s | 0.47t/s | 3.19t/s | |
Nova Micro |  | Proprietary | 130k | 28 | 0.53% | 0.36% | 0.05% | 0.14% | 0.09% | 0.8% | 0.7% | 0.08% | 0.71% | 0% | $0.06/1M | 0.14/1M | 0.04/1M | 335.6t/s | 304.7t/s | 314.6t/s | 362t/s | 383.6t/s | 0.32t/s | 0.28t/s | 0.31t/s | 0.33t/s | 0.38t/s | |
Gemini 1.5 Flash (May) |  | Proprietary | 1m | 28 | 0.57% | 0.32% | 0.04% | 0.2% | 0.18% | 0.72% | 0.55% | 0.09% | 0% | 1227% | $0.13/1M | 0.3/1M | 0.07/1M | 328.5t/s | 300.2t/s | 310.4t/s | 339.8t/s | 356.4t/s | 0.29t/s | 0.24t/s | 0.26t/s | 0.3t/s | 0.38t/s | |
Codestral (Jan '25) |  | Proprietary | 256k | 28 | 0.45% | 0.31% | 0.05% | 0.24% | 0.25% | 0.85% | 0.61% | 0.04% | 0% | 0% | $0.45/1M | 0.9/1M | 0.3/1M | 202.6t/s | 93.5t/s | 194.8t/s | 206t/s | 210.8t/s | 0.25t/s | 0.23t/s | 0.24t/s | 0.27t/s | 0.56t/s | |
Claude 3 Sonnet |  | Proprietary | 200k | 28 | 0.58% | 0.4% | 0.04% | 0.18% | 0.23% | 0.71% | 0.41% | 0.05% | 0% | 1201% | $6.0/1M | 15.0/1M | 3.0/1M | 61.5t/s | 55.3t/s | 59.3t/s | 62.6t/s | 63.9t/s | 0.52t/s | 0.48t/s | 0.5t/s | 0.6t/s | 1.17t/s | |
Qwen2.5 Coder 7B |  | Open | 131k | 27 | 0.47% | 0.34% | 0.05% | 0.13% | 0.15% | 0.9% | 0.66% | 0.05% | 0% | 0% | $0.03/1M | 0.06/1M | 0.02/1M | 204.6t/s | 201.4t/s | 202.9t/s | 218t/s | 219.9t/s | 0.48t/s | 0.45t/s | 0.47t/s | 0.52t/s | 0.66t/s | |
Phi-4 Multimodal |  | Open | 128k | 27 | 0.49% | 0.32% | 0.04% | 0.13% | 0.11% | 0.73% | 0.69% | 0.09% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 22.1t/s | 19.9t/s | 21.2t/s | 23.3t/s | 23.7t/s | 0.35t/s | 0.28t/s | 0.33t/s | 0.36t/s | 0.68t/s | |
Mistral Small (Sep '24) |  | Open | 33k | 27 | 0.53% | 0.38% | 0.04% | 0.14% | 0.16% | 0.81% | 0.56% | 0.06% | 0% | 0% | $0.3/1M | 0.6/1M | 0.2/1M | 67.1t/s | 11.1t/s | 52.4t/s | 81.2t/s | 99.5t/s | 0.33t/s | 0.28t/s | 0.3t/s | 0.49t/s | 0.96t/s | |
Llama 3 70B |  | Open | 8k | 27 | 0.57% | 0.38% | 0.04% | 0.2% | 0.19% | 0.79% | 0.48% | 0% | 0% | 1206% | $0.88/1M | 0.88/1M | 0.8/1M | 47.7t/s | 19t/s | 33t/s | 154t/s | 335.8t/s | 0.52t/s | 0.23t/s | 0.38t/s | 0.74t/s | 1.76t/s | |
Mistral Large (Feb '24) |  | Proprietary | 33k | 26 | 0.52% | 0.35% | 0.03% | 0.18% | 0.21% | 0.71% | 0.53% | 0% | 0% | 1157% | $6.0/1M | 12.0/1M | 4.0/1M | 30.1t/s | 22.2t/s | 29.2t/s | 30.9t/s | 31.5t/s | 0.46t/s | 0.42t/s | 0.45t/s | 0.5t/s | 1.37t/s | |
Phi-4 Mini |  | Open | 128k | 26 | 0.47% | 0.33% | 0.04% | 0.13% | 0.11% | 0.74% | 0.7% | 0.03% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 56.7t/s | 52.5t/s | 55t/s | 57.5t/s | 59t/s | 0.34t/s | 0.31t/s | 0.33t/s | 0.37t/s | 0.49t/s | |
Mixtral 8x22B |  | Open | 65k | 26 | 0.54% | 0.33% | 0.04% | 0.15% | 0.19% | 0.72% | 0.55% | 0% | 0% | 1148% | $3.0/1M | 6.0/1M | 2.0/1M | 55.8t/s | 35.1t/s | 49.2t/s | 66.4t/s | 80.4t/s | 0.35t/s | 0.31t/s | 0.32t/s | 0.4t/s | 0.79t/s | |
Phi-3 Medium 14B |  | Open | 128k | 25 | 0.54% | 0.33% | 0.05% | 0.15% | 0.12% | 0% | 0.46% | 0.01% | 0% | 1123% | $0.3/1M | 0.68/1M | 0.17/1M | 51.9t/s | 34.4t/s | 50.6t/s | 53.1t/s | 54.6t/s | 0.43t/s | 0.39t/s | 0.42t/s | 0.46t/s | 0.52t/s | |
Llama 3.1 8B |  | Open | 128k | 24 | 0.48% | 0.26% | 0.05% | 0.12% | 0.13% | 0.67% | 0.52% | 0.08% | 0.61% | 1172% | $0.1/1M | 0.1/1M | 0.1/1M | 155.1t/s | 53.3t/s | 74.1t/s | 465.8t/s | 2039.00t/s | 0.36t/s | 0.16t/s | 0.23t/s | 0.5t/s | 1.13t/s | |
Claude 2.1 |  | Proprietary | 200k | 24 | 0.5% | 0.32% | 0.04% | 0.2% | 0.18% | 0.16% | 0.37% | 0.03% | 0% | 1118% | $12.0/1M | 24.0/1M | 8.0/1M | 13.8t/s | 13.5t/s | 13.7t/s | 14.1t/s | 14.4t/s | 0.85t/s | 0.79t/s | 0.81t/s | 0.87t/s | 1.2t/s | |
Gemma 3 4B |  | Open | 128k | 24 | 0.42% | 0.29% | 0.05% | 0.07% | 0.06% | 0.72% | 0.77% | 0.05% | 0% | 0% | $0.03/1M | 0.04/1M | 0.02/1M | 95.8t/s | 50.5t/s | 86.6t/s | 101.5t/s | 108.8t/s | 0.3t/s | 0.25t/s | 0.27t/s | 0.34t/s | 0.64t/s | |
Pixtral 12B |  | Open | 128k | 23 | 0.47% | 0.34% | 0.05% | 0.12% | 0.14% | 0.78% | 0.46% | 0% | 0% | 0% | $0.15/1M | 0.15/1M | 0.15/1M | 110.2t/s | 101.5t/s | 106.4t/s | 112.7t/s | 117.1t/s | 0.3t/s | 0.26t/s | 0.27t/s | 0.35t/s | 0.5t/s | |
Mistral Medium |  | Proprietary | 33k | 23 | 0.49% | 0.35% | 0.03% | 0.1% | 0.12% | 0% | 0.41% | 0.04% | 0% | 1148% | $4.09/1M | 8.1/1M | 2.75/1M | 41.6t/s | 22.5t/s | 40.3t/s | 42.5t/s | 43.5t/s | 0.38t/s | 0.34t/s | 0.36t/s | 0.44t/s | 0.99t/s | |
Mistral Small (Feb '24) |  | Proprietary | 33k | 23 | 0.42% | 0.3% | 0.04% | 0.11% | 0.13% | 0.79% | 0.56% | 0.01% | 0% | 0% | $1.5/1M | 3.0/1M | 1.0/1M | 150.4t/s | 62.4t/s | 116.2t/s | 161.7t/s | 172.3t/s | 0.29t/s | 0.24t/s | 0.26t/s | 0.35t/s | 2.19t/s | |
LFM 40B |  | Proprietary | 32k | 22 | 0.43% | 0.33% | 0.05% | 0.1% | 0.07% | 0.51% | 0.48% | 0.02% | 0% | 0% | $0.15/1M | 0.15/1M | 0.15/1M | 171.6t/s | 154.1t/s | 165.4t/s | 174.6t/s | 177.1t/s | 0.46t/s | 0.15t/s | 0.25t/s | 0.49t/s | 0.58t/s | |
Phi-3 Mini |  | Open | 4k | 22 | 0.44% | 0.32% | 0.04% | 0.12% | 0.09% | 0.25% | 0.46% | 0.04% | 0% | 1037% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Gemma 2 9B |  | Open | 8k | 22 | 0.5% | 0.31% | 0.04% | 0.13% | 0.01% | 0.65% | 0.52% | 0% | 0% | 1190% | $0.12/1M | 0.15/1M | 0.12/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Ministral 8B |  | Open | 128k | 22 | 0.39% | 0.28% | 0.05% | 0.11% | 0.12% | 0.77% | 0.57% | 0.04% | 0% | 1183% | $0.1/1M | 0.1/1M | 0.1/1M | 142.3t/s | 79.8t/s | 137.3t/s | 145.7t/s | 147.6t/s | 0.28t/s | 0.26t/s | 0.27t/s | 0.29t/s | 0.67t/s | |
Llama 3 8B |  | Open | 8k | 21 | 0.41% | 0.3% | 0.05% | 0.1% | 0.12% | 0.71% | 0.5% | 0% | 0% | 1152% | $0.09/1M | 0.14/1M | 0.06/1M | 92t/s | 26.2t/s | 73.8t/s | 123.5t/s | 1350.80t/s | 0.36t/s | 0.21t/s | 0.3t/s | 0.5t/s | 1.59t/s | |
Command-R+ |  | Open | 128k | 21 | 0.43% | 0.34% | 0.05% | 0.11% | 0.12% | 0.63% | 0.4% | 0% | 0% | 1215% | $4.38/1M | 10.0/1M | 2.5/1M | 53.5t/s | 50.2t/s | 51.8t/s | 54t/s | 55.3t/s | 0.27t/s | 0.25t/s | 0.26t/s | 0.29t/s | 0.32t/s | |
Gemini 1.0 Pro |  | Proprietary | 33k | 21 | 0.43% | 0.28% | 0.05% | 0.12% | 0.12% | 0.02% | 0.4% | 0.01% | 0% | 1111% | $0.75/1M | 1.5/1M | 0.5/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Mistral NeMo |  | Open | 128k | 20 | 0.4% | 0.31% | 0.04% | 0.06% | 0.1% | 0.65% | 0.4% | 0% | 0% | 0% | $0.15/1M | 0.15/1M | 0.15/1M | 140.7t/s | 75.8t/s | 134.6t/s | 145.1t/s | 148.8t/s | 0.29t/s | 0.26t/s | 0.27t/s | 0.33t/s | 0.48t/s | |
Ministral 3B |  | Proprietary | 128k | 20 | 0.34% | 0.26% | 0.06% | 0.07% | 0.09% | 0.74% | 0.54% | 0% | 0% | 0% | $0.04/1M | 0.04/1M | 0.04/1M | 225.5t/s | 132.8t/s | 220.8t/s | 231.6t/s | 239.7t/s | 0.27t/s | 0.24t/s | 0.25t/s | 0.35t/s | 0.75t/s | |
DBRX |  | Open | 33k | 20 | 0.4% | 0.33% | 0.07% | 0.09% | 0.12% | 0.67% | 0.28% | 0.03% | 0% | 1103% | $1.13/1M | 2.25/1M | 0.75/1M | 69.6t/s | 51.3t/s | 69.2t/s | 70.1t/s | 70.7t/s | 0.55t/s | 0.44t/s | 0.46t/s | 0.73t/s | 0.97t/s | |
Command-R+ (Apr '24) |  | Open | 128k | 20 | 0.43% | 0.32% | 0.05% | 0.12% | 0.12% | 0.64% | 0.28% | 0.01% | 0% | 1190% | $6.0/1M | 15.0/1M | 3.0/1M | 67.8t/s | 58.3t/s | 61.9t/s | 72t/s | 77.6t/s | 0.23t/s | 0.21t/s | 0.22t/s | 0.24t/s | 0.25t/s | |
Llama 2 Chat 13B |  | Open | 4k | 20 | 0.41% | 0.32% | 0.05% | 0.1% | 0.12% | 0% | 0.33% | 0.02% | 0% | 1063% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Aya Expanse 32B |  | Open | 128k | 20 | 0.38% | 0.23% | 0.05% | 0.14% | 0.15% | 0.68% | 0.45% | 0% | 0.65% | 1207% | $0.75/1M | 1.5/1M | 0.5/1M | 122t/s | 118.1t/s | 120.6t/s | 127t/s | 128.8t/s | 0.16t/s | 0.14t/s | 0.15t/s | 0.17t/s | 0.22t/s | |
Codestral (May '24) |  | Open | 33k | 20 | 0.33% | 0.26% | 0.05% | 0.21% | 0.22% | 0.8% | 0.35% | 0% | 0% | 0% | $0.3/1M | 0.6/1M | 0.2/1M | 106.5t/s | 46.6t/s | 103.1t/s | 108.4t/s | 110t/s | 0.3t/s | 0.28t/s | 0.29t/s | 0.35t/s | 0.87t/s | |
Llama 3.2 3B |  | Open | 128k | 20 | 0.35% | 0.26% | 0.05% | 0.08% | 0.05% | 0.56% | 0.49% | 0.07% | 0% | 1103% | $0.05/1M | 0.06/1M | 0.04/1M | 160.4t/s | 62.1t/s | 87.2t/s | 244.1t/s | 1584.50t/s | 0.45t/s | 0.18t/s | 0.28t/s | 0.53t/s | 1.11t/s | |
DeepSeek R1 Distill Qwen 1.5B |  | Open | 128k | 19 | 0.27% | 0.1% | 0.03% | 0.07% | 0.07% | 0.45% | 0.69% | 0.18% | 0% | 0% | $0.18/1M | 0.18/1M | 0.18/1M | 315.9t/s | 259.4t/s | 275.1t/s | 373.7t/s | 390.3t/s | 0.25t/s | 0.17t/s | 0.22t/s | 0.27t/s | 0.35t/s | |
Jamba 1.6 Mini |  | Open | 256k | 18 | 0.37% | 0.3% | 0.05% | 0.07% | 0.1% | 0.43% | 0.26% | 0.03% | 0% | 0% | $0.25/1M | 0.4/1M | 0.2/1M | 198.9t/s | 183.2t/s | 196.8t/s | 203.1t/s | 205t/s | 0.37t/s | 0.19t/s | 0.3t/s | 0.4t/s | 0.8t/s | |
Jamba 1.5 Mini |  | Open | 256k | 18 | 0.37% | 0.3% | 0.05% | 0.06% | 0.08% | 0.63% | 0.36% | 0.01% | 0% | 1176% | $0.25/1M | 0.4/1M | 0.2/1M | 167.4t/s | 140.4t/s | 160.2t/s | 184.8t/s | 197.8t/s | 0.34t/s | 0.16t/s | 0.22t/s | 0.38t/s | 0.58t/s | |
Mixtral 8x7B |  | Open | 33k | 17 | 0.39% | 0.29% | 0.05% | 0.07% | 0.03% | 0.01% | 0.3% | 0% | 0% | 1114% | $0.7/1M | 0.7/1M | 0.7/1M | 92t/s | 28.8t/s | 77.8t/s | 99.9t/s | 102.4t/s | 0.3t/s | 0.28t/s | 0.29t/s | 0.35t/s | 0.8t/s | |
Aya Expanse 8B |  | Open | 8k | 16 | 0.31% | 0.25% | 0.05% | 0.07% | 0.08% | 0.44% | 0.32% | 0% | 0.49% | 0% | $0.75/1M | 1.5/1M | 0.5/1M | 168.7t/s | 162.2t/s | 166.6t/s | 172.5t/s | 176.6t/s | 0.12t/s | 0.1t/s | 0.11t/s | 0.12t/s | 0.18t/s | |
DeepHermes 3 - Llama-3.1 8B |  | Open | 128k | 16 | 0.37% | 0.27% | 0.04% | 0.09% | 0.09% | 0.54% | 0.22% | 0% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Command-R |  | Open | 128k | 15 | 0.34% | 0.29% | 0.05% | 0.04% | 0.09% | 0.42% | 0.15% | 0% | 0% | 1179% | $0.26/1M | 0.6/1M | 0.15/1M | 86.9t/s | 80.6t/s | 84.4t/s | 88.3t/s | 89.6t/s | 0.19t/s | 0.18t/s | 0.18t/s | 0.21t/s | 0.22t/s | |
Command-R (Mar '24) |  | Open | 128k | 15 | 0.34% | 0.28% | 0.05% | 0.05% | 0.06% | 0.4% | 0.16% | 0.01% | 0% | 1149% | $0.75/1M | 1.5/1M | 0.5/1M | 179.5t/s | 166.4t/s | 176.6t/s | 182t/s | 185.2t/s | 0.14t/s | 0.13t/s | 0.14t/s | 0.16t/s | 0.18t/s | |
Codestral-Mamba |  | Open | 256k | 14 | 0.21% | 0.21% | 0.05% | 0.13% | 0.11% | 0.8% | 0.24% | 0% | 0% | 0% | $0.25/1M | 0.25/1M | 0.25/1M | 93.9t/s | 54.6t/s | 89.9t/s | 95.5t/s | 97.5t/s | 0.44t/s | 0.41t/s | 0.42t/s | 0.5t/s | 0.86t/s | |
Llama 3.2 1B |  | Open | 128k | 10 | 0.2% | 0.2% | 0.05% | 0.02% | 0.02% | 0.4% | 0.14% | 0% | 0% | 1054% | $0.03/1M | 0.05/1M | 0.03/1M | 181.6t/s | 95.4t/s | 118.5t/s | 275.8t/s | 2632.70t/s | 0.44t/s | 0.17t/s | 0.2t/s | 0.48t/s | 2.51t/s | |
Mistral 7B |  | Open | 8k | 10 | 0.25% | 0.18% | 0.04% | 0.05% | 0.02% | 0.4% | 0.12% | 0% | 0% | 1008% | $0.25/1M | 0.25/1M | 0.25/1M | 113.5t/s | 51.6t/s | 110.6t/s | 116.9t/s | 120.2t/s | 0.28t/s | 0.26t/s | 0.27t/s | 0.32t/s | 0.65t/s | |
Llama 2 Chat 7B |  | Open | 4k | 8 | 0.16% | 0.23% | 0.06% | 0% | 0% | 0% | 0.06% | 0% | 0% | 1037% | $0.1/1M | 0.25/1M | 0.05/1M | 132t/s | 108t/s | 127.8t/s | 134.1t/s | 135t/s | 0.43t/s | 0.38t/s | 0.41t/s | 8.08t/s | 20.73t/s | |
Reka Flash (Feb '24) |  | Proprietary | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.61% | 0.33% | 0% | 0% | 1165% | $0.35/1M | 0.8/1M | 0.2/1M | 46.1t/s | 45t/s | 45.8t/s | 46.3t/s | 46.4t/s | 0.93t/s | 0.86t/s | 0.88t/s | 0.97t/s | 1.03t/s | |
Claude 3.5 Sonnet (June) |  | Proprietary | 200k | 0 | 0.75% | 0.56% | 0.04% | 0% | 0.32% | 0.9% | 0.7% | 0.1% | 0% | 1268% | $6.0/1M | 15.0/1M | 3.0/1M | 80.3t/s | 69.8t/s | 78.5t/s | 82.6t/s | 85.7t/s | 0.62t/s | 0.5t/s | 0.54t/s | 0.71t/s | 1.17t/s | |
Llama 2 Chat 70B |  | Open | 4k | 0 | 0.41% | 0.33% | 0.05% | 0.1% | 0% | 0% | 0.32% | 0% | 0% | 1093% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
R1 1776 |  | Open | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0.95% | 0% | 0% | 0% | $3.5/1M | 8.0/1M | 2.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
o3 |  | Proprietary | 128k | 0 | 0% | 0.83% | 0.18% | 0.53% | 0% | 0% | 0.99% | 0.9% | 0% | 0% | $17.5/1M | 40.0/1M | 10.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Reka Core |  | Proprietary | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.73% | 0.56% | 0% | 0% | 1199% | $2.0/1M | 2.0/1M | 2.0/1M | 27.8t/s | 26.6t/s | 27.3t/s | 28.1t/s | 28.4t/s | 0.96t/s | 0.87t/s | 0.91t/s | 1.02t/s | 1.16t/s | |
Gemini 2.5 Flash Preview |  | Proprietary | 1m | 0 | 0% | 0.7% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $0.26/1M | 0.6/1M | 0.15/1M | 214.9t/s | 176.4t/s | 195.8t/s | 229.5t/s | 375.8t/s | 10.76t/s | 6.08t/s | 9.77t/s | 11.45t/s | 14.38t/s | |
GPT-4 |  | Proprietary | 8k | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 1186% | $37.5/1M | 60.0/1M | 30.0/1M | 32.1t/s | 21.6t/s | 26.2t/s | 34.7t/s | 37.7t/s | 0.65t/s | 0.55t/s | 0.59t/s | 0.7t/s | 0.89t/s | |
Sonar Reasoning Pro |  | Proprietary | 127k | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0.96% | 0.79% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
GPT-4o (Aug '24) |  | Proprietary | 128k | 0 | 0% | 0.52% | 0.03% | 0.32% | 0% | 0.93% | 0.8% | 0.12% | 0.87% | 1337% | $4.38/1M | 10.0/1M | 2.5/1M | 126.1t/s | 76.9t/s | 95t/s | 144.4t/s | 163.3t/s | 0.52t/s | 0.44t/s | 0.48t/s | 0.54t/s | 0.58t/s | |
Reka Flash |  | Proprietary | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.74% | 0.53% | 0% | 0% | 0% | $0.35/1M | 0.8/1M | 0.2/1M | 46.2t/s | 45.4t/s | 46t/s | 46.2t/s | 46.4t/s | 0.95t/s | 0.86t/s | 0.89t/s | 1t/s | 1.07t/s | |
Gemini 2.0 Flash Thinking exp. (Dec '24) |  | Proprietary | 2m | 0 | 0% | 0% | 0% | 0% | 0% | 0.94% | 0.48% | 0% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
GPT-4o (ChatGPT) |  | Proprietary | 128k | 0 | 0.77% | 0.51% | 0.04% | 0% | 0.33% | 0.94% | 0.8% | 0.1% | 0% | 0% | $7.5/1M | 15.0/1M | 5.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
DeepSeek Coder V2 Lite |  | Open | 128k | 0 | 0.43% | 0.32% | 0.05% | 0.16% | 0.14% | 0% | 0% | 0% | 0% | 0% | $0.09/1M | 0.18/1M | 0.06/1M | 115.7t/s | 105t/s | 112.5t/s | 118.5t/s | 121.5t/s | 0.6t/s | 0.55t/s | 0.58t/s | 0.67t/s | 0.81t/s | |
o1-preview |  | Proprietary | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.96% | 0.92% | 0% | 0.86% | 1334% | $26.25/1M | 60.0/1M | 15.0/1M | 162.9t/s | 77t/s | 142.5t/s | 171t/s | 180.2t/s | 18.56t/s | 12.6t/s | 15.57t/s | 21.4t/s | 27.82t/s | |
Jamba Instruct |  | Proprietary | 256k | 0 | 0.34% | 0.27% | 0.05% | 0.05% | 0.08% | 0% | 0.24% | 0% | 0% | 0% | $0.55/1M | 0.7/1M | 0.5/1M | 164.6t/s | 142t/s | 155.6t/s | 175.9t/s | 188.8t/s | 0.33t/s | 0.17t/s | 0.25t/s | 0.41t/s | 0.55t/s | |
Llama 3.1 Nemotron Ultra 253B Reasoning |  | Open | 128k | 0 | 0% | 0.73% | 0.08% | 0.64% | 0% | 0% | 0.96% | 0.75% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Gemini 2.5 Flash Preview (Reasoning) |  | Proprietary | 1m | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $0.99/1M | 3.5/1M | 0.15/1M | 203.6t/s | 145.3t/s | 178.3t/s | 299.4t/s | 389.4t/s | 11.44t/s | 6.67t/s | 10.27t/s | 12.07t/s | 15.76t/s | |
GPT-4 Turbo |  | Proprietary | 128k | 0 | 0.69% | 0% | 0.03% | 0.29% | 0.32% | 0.92% | 0.74% | 0.15% | 0% | 1256% | $15.0/1M | 30.0/1M | 10.0/1M | 46t/s | 21.5t/s | 35.9t/s | 51.1t/s | 59.7t/s | 0.73t/s | 0.56t/s | 0.61t/s | 0.89t/s | 1.52t/s | |
DeepSeek LLM 67B (V1) |  | Open | 4k | 0 | 0% | 0% | 0% | 0% | 0% | 0.75% | 0% | 0% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
OpenChat 3.5 |  | Open | 8k | 0 | 0.31% | 0.23% | 0.05% | 0.12% | 0% | 0.68% | 0.31% | 0% | 0% | 1076% | $0.06/1M | 0.06/1M | 0.06/1M | 47.4t/s | 30.4t/s | 40.7t/s | 52.5t/s | 68t/s | 0.24t/s | 0.15t/s | 0.2t/s | 0.32t/s | 0.46t/s | |
Gemma 3 1B |  | Open | 32k | 0 | 0.1% | 0.19% | 0% | 0.01% | 0% | 0.32% | 0.14% | 0.01% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Qwen Chat 72B |  | Open | 34k | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $1.0/1M | 1.0/1M | 1.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
DeepSeek-Coder-V2 |  | Open | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.87% | 0.74% | 0% | 0% | 1178% | $0.17/1M | 0.28/1M | 0.14/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
DeepSeek-V2 |  | Open | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.87% | 0% | 0% | 0% | 1220% | $0.17/1M | 0.28/1M | 0.14/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Solar Mini |  | Open | 4k | 0 | 0% | 0% | 0% | 0% | 0% | 0.59% | 0.33% | 0% | 0% | 1062% | $0.15/1M | 0.15/1M | 0.15/1M | 49.9t/s | 42.7t/s | 46.7t/s | 86.6t/s | 89.2t/s | 1.07t/s | 0.99t/s | 1.04t/s | 1.12t/s | 1.24t/s | |
Gemma 2 27B |  | Open | 8k | 0 | 0.57% | 0.36% | 0.04% | 0% | 0.13% | 0.76% | 0.54% | 0.09% | 0% | 1219% | $0.26/1M | 0.51/1M | 0.17/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Qwen1.5 Chat 110B |  | Open | 32k | 0 | 0% | 0.29% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 23.7t/s | 23.6t/s | 23.6t/s | 23.8t/s | 23.8t/s | 1.61t/s | 1.48t/s | 1.52t/s | 1.68t/s | 1.74t/s | |
DeepSeek-V2.5 (Dec '24) |  | Open | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.88% | 0.76% | 0% | 0% | 0% | $0.17/1M | 0.28/1M | 0.14/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
DeepSeek-V2.5 |  | Open | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.87% | 0% | 0% | 0% | 1258% | $0.17/1M | 0.28/1M | 0.14/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Grok 3 mini |  | Proprietary | 1m | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
GPT-4o mini Realtime (Dec '24) |  | Proprietary | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Llama 3.2 11B (Vision) |  | Open | 128k | 0 | 0.46% | 0.22% | 0.05% | 0.11% | 0% | 0.69% | 0.52% | 0.09% | 0% | 0% | $0.16/1M | 0.16/1M | 0.16/1M | 103.6t/s | 41t/s | 82.7t/s | 128.2t/s | 144.1t/s | 0.35t/s | 0.13t/s | 0.24t/s | 0.44t/s | 0.51t/s | |
Arctic |  | Open | 4k | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 1090% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
Mistral Saba |  | Proprietary | 32k | 0 | 0.61% | 0.42% | 0.04% | 0% | 0.24% | 0.85% | 0.68% | 0.13% | 0% | 0% | $0.3/1M | 0.6/1M | 0.2/1M | 91.6t/s | 47.5t/s | 76.5t/s | 95.3t/s | 99.3t/s | 0.32t/s | 0.29t/s | 0.3t/s | 0.38t/s | 0.72t/s | |
Claude 2.0 |  | Proprietary | 100k | 0 | 0.49% | 0.34% | 0% | 0.17% | 0.19% | 0% | 0% | 0% | 0% | 1132% | $12.0/1M | 24.0/1M | 8.0/1M | 30.9t/s | 29.8t/s | 30.4t/s | 31.6t/s | 33.3t/s | 0.83t/s | 0.79t/s | 0.8t/s | 0.87t/s | 0.95t/s | |
Grok 3 Reasoning Beta |  | Proprietary | 1m | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
GPT-4o Realtime (Dec '24) |  | Proprietary | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $0.0/1M | 0.0/1M | 0.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |
GPT-4.5 (Preview) |  | Proprietary | 128k | 0 | 0% | 0.71% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $93.75/1M | 150.0/1M | 75.0/1M | 74.9t/s | 51.3t/s | 61.3t/s | 84.2t/s | 91.4t/s | 0.93t/s | 0.84t/s | 0.9t/s | 1.08t/s | 1.34t/s | |
Reka Edge |  | Proprietary | 128k | 0 | 0% | 0% | 0% | 0% | 0% | 0.41% | 0.22% | 0% | 0% | 0% | $0.1/1M | 0.1/1M | 0.1/1M | 84.6t/s | 81.7t/s | 82.9t/s | 86.6t/s | 87.9t/s | 0.92t/s | 0.83t/s | 0.86t/s | 0.97t/s | 1.05t/s | |
Claude 3 Haiku |  | Proprietary | 200k | 0 | 0% | 0% | 0% | 0.15% | 0.19% | 0.76% | 0.39% | 0.01% | 0.68% | 1179% | $0.5/1M | 1.25/1M | 0.25/1M | 140.9t/s | 116.3t/s | 134.7t/s | 147.9t/s | 157t/s | 0.47t/s | 0.36t/s | 0.4t/s | 0.64t/s | 0.98t/s | |
Gemini 2.0 Flash-Lite (Preview) |  | Proprietary | 1m | 0 | 0% | 0.54% | 0.04% | 0.18% | 0.25% | 0.9% | 0.87% | 0.3% | 0% | 0% | $0.13/1M | 0.3/1M | 0.07/1M | 198.2t/s | 177t/s | 189.9t/s | 209.6t/s | 216.6t/s | 0.29t/s | 0.24t/s | 0.26t/s | 0.31t/s | 0.36t/s | |
Sonar Reasoning |  | Proprietary | 127k | 0 | 0% | 0.62% | 0% | 0% | 0% | 0% | 0.92% | 0.77% | 0% | 0% | $2.0/1M | 5.0/1M | 1.0/1M | 78t/s | 67.9t/s | 75.8t/s | 80.9t/s | 85.9t/s | 1.8t/s | 1.43t/s | 1.54t/s | 2.05t/s | 2.46t/s | |
o1-pro |  | Proprietary | 200k | 0 | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | $262.5/1M | 600.0/1M | 150.0/1M | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | 0t/s | |