| Model | Context | Input | Output | Throughput | Latency | Providers | Agents | Coding | Reasoning | Knowledge | Math | Multilingual | Multimodal | Instruction following |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Gemini 2.5 Flash Live | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MiniMax M2.1 Lightning | — | — | — | — | — | +3 | — | — | — | — | — | — | — | — |
| Gemini 3 Pro DeepSearch | — | — | — | — | — | +1 | — | — | — | — | — | — | — | — |
| DeepSeek V3.2 Reasoner | — | — | — | — | — | +1 | — | — | — | — | — | — | — | — |
| Qwen3.5 Omni Plus | — | — | — | 36 t/s | 2.15s | +19 | — | 40.5 | 48.3 | — | — | — | — | — |
| MiniMax Voice Clone | — | — | — | — | — | +7 | — | — | — | — | — | — | — | — |
| DeepSeek Vision Reasoner | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Gemini Robotics ER 1.5 | — | — | — | — | — | +17 | — | — | — | — | — | — | — | — |
| GLM-4 FlashX | — | — | — | 60 t/s | 0.41s | +8 | — | — | — | — | — | — | — | — |
| Qwen3.5 Omni Flash | — | — | — | — | — | +19 | — | 25.5 | 40.7 | — | — | — | — | — |
| Claude 2.0 | — | — | — | — | — | +8 | — | 18.3 | 41.5 | — | — | — | — | — |
| Gemini 3.0 Pro | — | — | — | — | — | +1 | — | — | — | — | — | — | — | — |
| MiniMax File Upload | — | — | — | — | — | +7 | — | — | — | — | — | — | — | — |
| Gemini 2.5 Flash DeepSearch | — | — | — | — | — | +2 | — | — | — | — | — | — | — | — |
| DeepSeek R1 Turbo | — | — | — | — | — | +1 | — | — | — | — | — | — | — | — |
| Gemini 1.5 Pro 002 | — | — | — | — | — | +2 | — | — | — | — | — | — | — | — |
| GPT-4o Study | — | — | — | — | — | +4 | — | — | — | — | — | — | — | — |
| MiniMax Voice Design | — | — | — | — | — | +7 | — | — | — | — | — | — | — | — |
| Gemini 3.0 Pro Image | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DeepSeek V3 Turbo | — | — | — | 28 t/s | 1.31s | — | — | — | — | — | — | — | — | |
| Gemini 2.5 Pro DeepSearch | — | — | — | — | — | +2 | — | — | — | — | — | — | — | — |
| Qwen3.5 Plus Image Edit | — | — | — | — | — | +1 | — | — | — | — | — | — | — | — |
| MiniMax File | — | — | — | — | — | +1 | — | — | — | — | — | — | — | — |
| DeepSeek V3.1 | — | — | — | 103 t/s | 1.78s | +110 | 37.4 | 39.6 | 43.7 | 44.1 | — | — | 100 | 41.5 |
| DeepSeek V2.5 | — | — | — | 16 t/s | 0.96s | +3 | — | — | — | — | 76.3 | — | — | — |
| Meta Llama 3.1 Instruct Turbo | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Meta Llama 3.3 Instruct Turbo | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Meta Llama 3.1 Instruct | — | — | — | — | — | +3 | — | — | — | — | — | — | — | — |
| DeepSeek R1 Distill Qwen 1.5B | — | — | — | 126 t/s | 4.28s | +19 | — | 6.8 | 13.3 | — | 43.2 | — | — | — |
| Usdcode Llama 3.1 Instruct | — | — | — | — | — | +5 | — | — | — | — | — | — | — | — |
| Dracarys Llama 3.1 Instruct | — | — | — | 18 t/s | 0.59s | +22 | — | — | — | — | — | — | — | — |
| Meta Llama 3.3 Instruct | — | — | — | 48 t/s | 0.93s | — | — | — | — | — | — | — | — | |
| Llama 3.3 | — | — | — | 985 t/s | 0.45s | +5 | — | — | — | — | — | — | — | — |
| Llama 3.1 | — | — | — | — | — | +18 | 19 | 18.3 | 24.3 | 27.5 | — | — | — | 39 |
| LFM 2.5 1.2B Instruct | 32.8K | — | — | — | — | +15 | — | — | — | — | — | — | — | — |
| Qwen1.8B | — | — | — | — | — | +13 | — | — | — | — | — | — | — | — |
| Qwen1.8B Long Context | — | — | — | — | — | +13 | — | — | — | — | — | — | — | — |
| DeepSeek V4 | — | — | — | 47 t/s | 2.49s | +8 | — | — | — | — | — | — | — | — |
| Claude Opus 4.7 Max | — | — | — | — | — | +20 | 73.3 | 66.8 | 61 | 54.2 | 43.8 | — | 79.1 | 58.6 |
| MiMo-V2-TTS | — | — | — | — | — | +32 | — | — | — | — | — | — | — | — |
| Kimi K2.6 | 262.1K | — | — | 53 t/s | 18.41s | +104 | 61.9 | 61.4 | 51.2 | 48.6 | 91.7 | — | 86.2 | 76 |
| MiMo-V2.5 | 1.0M | — | — | 82 t/s | 3.72s | +89 | 43.1 | 55 | 55.1 | — | — | — | 86.7 | — |
| MiMo-V2.5-Pro | 1.0M | — | — | 45 t/s | 5.93s | +93 | 59.1 | 54.9 | 49.4 | 35.3 | — | — | 100 | 79.9 |
| Qwen3 Instant | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MiMo-V2.5-TTS | — | — | — | — | — | +35 | — | — | — | — | — | — | — | — |
| MiMo-V2.5-TTS-VoiceClone | — | — | — | — | — | +35 | — | — | — | — | — | — | — | — |
| MiMo-V2.5-TTS-VoiceDesign | — | — | — | — | — | +32 | — | — | — | — | — | — | — | — |
| DeepSeek V4 Flash | 1.0M | — | — | 69 t/s | 7.31s | +180 | 60.4 | 63.2 | 58.6 | 62 | 75.5 | — | 100 | 79.2 |
| DeepSeek V4 Pro | 1.0M | — | — | 43 t/s | 8.37s | +172 | 57.6 | 66 | 62.3 | 67.3 | 78.4 | — | 100 | 76.5 |
| Qwen3.5 Plus | 1M | — | — | 51 t/s | 13.81s | +102 | — | 15.7 | — | — | — | — | — | — |
| Gemini Flash | 1.0M | — | — | — | — | +36 | — | — | — | — | — | — | — | — |
| Gemini Pro | 1.0M | — | — | — | — | +32 | — | — | — | — | — | — | — | — |
| Gemini 3.1 Flash Lite | 1.0M | — | — | 183 t/s | 6.98s | +101 | 32.5 | 33.6 | 33.2 | 48.3 | — | — | 74.4 | 77.2 |
| Claude Opus 4.7 | 1M | — | — | 47 t/s | 4.89s | +169 | 53.4 | 58.6 | 50.8 | 45.3 | — | — | 88.2 | 43.6 |
| Claude Opus 4.8 | 1M | — | — | 255 t/s | 2.22s | +122 | 70.9 | 63.5 | 56.6 | 55.8 | 96.7 | 87.6 | 84.9 | 62.2 |
| Claude Fable 5 | 1M | — | — | 72 t/s | 2.83s | +89 | 89.1 | 73.4 | 73 | 59 | 97.6 | 87.3 | 73.9 | — |
| Kimi K2.7 | 262.1K | — | — | — | — | +30 | — | — | — | — | — | — | — | — |
| GLM-5.2 | 1.0M | — | — | 61 t/s | 6.89s | +102 | 61.5 | 59 | 55.5 | 46.2 | 94.3 | — | 100 | 73.3 |
| Grok 4.5 | 500K | — | — | — | — | +1 | 83.3 | 69.4 | 66.7 | — | — | — | — | — |
| Nemotron 3 Ultra | 1M | — | — | — | — | +9 | — | 39.9 | 56.7 | — | — | — | — | — |
