Output token price per 1M tokens.
Output Price measures the listed cost of generated text. Output tokens are often priced higher than input tokens, so this metric can dominate workloads with long answers or generated artifacts.
Test type: Provider output-token price comparison. Lower values rank better.
357 models have this metric.
Current leader: Gemma 3n E4B Instruct
Project links
Prices come from Artificial Analysis pricing data in the committed snapshot.
Top models ranked by Output $.
| Rank | Model | Creator | Value | Speed | Blended Price |
|---|---|---|---|---|---|
| #1 | Gemma 3n E4B Instruct | $0.040/M | 50 tok/s | $0.025/M | |
| #2 |
| Meta |
| $0.050/M |
| 92.9 tok/s |
| $0.050/M |
| #3 | Qwen3.5 0.8B (Non-reasoning) | Alibaba | $0.050/M | 88 tok/s | $0.020/M |
| #4 | Qwen3.5 0.8B (Reasoning) | Alibaba | $0.050/M | n/a | $0.020/M |
| #5 | Gemma 3 4B Instruct | $0.080/M | n/a | $0.050/M |
| #6 | Granite 4.1 8B | IBM | $0.100/M | 134.2 tok/s | $0.063/M |
| #7 | Llama 3.1 Instruct 8B | Meta | $0.100/M | 201.5 tok/s | $0.100/M |
| #8 | Ministral 3 3B | Mistral | $0.100/M | 174.3 tok/s | $0.100/M |
| #9 | Qwen3.5 2B (Non-reasoning) | Alibaba | $0.100/M | 318.9 tok/s | $0.040/M |
| #10 | Qwen3.5 2B (Reasoning) | Alibaba | $0.100/M | n/a | $0.040/M |
| #11 | Sarvam 30B (high) | Sarvam | $0.110/M | 147.9 tok/s | $0.047/M |
| #12 | LFM2 24B A2B | Liquid AI | $0.120/M | 127.6 tok/s | $0.052/M |
| #13 | Nova Micro | Amazon | $0.140/M | 284.2 tok/s | $0.061/M |
| #14 | Llama 3 Instruct 8B | Meta | $0.145/M | 88.3 tok/s | $0.070/M |
| #15 | Llama 3.2 Instruct 3B | Meta | $0.150/M | 52.3 tok/s | $0.150/M |
| #16 | Ministral 3 8B | Mistral | $0.150/M | 103.6 tok/s | $0.150/M |
| #17 | Qwen3.5 4B (Non-reasoning) | Alibaba | $0.150/M | 208.9 tok/s | $0.060/M |
| #18 | Qwen3.5 4B (Reasoning) | Alibaba | $0.150/M | 195.8 tok/s | $0.060/M |
| #19 | Qwen3.5 9B (Reasoning) | Alibaba | $0.150/M | 69.4 tok/s | $0.113/M |
| #20 | Solar Mini | Upstage | $0.150/M | 75.9 tok/s | $0.150/M |
| #21 | NVIDIA Nemotron Nano 9B V2 (Reasoning) | NVIDIA | $0.160/M | 118 tok/s | $0.070/M |
| #22 | Sarvam 105B (high) | Sarvam | $0.170/M | 100.7 tok/s | $0.074/M |
| #23 | Mistral Small 3 | Mistral | $0.190/M | 153.7 tok/s | $0.104/M |
| #24 | NVIDIA Nemotron Nano 9B V2 (Non-reasoning) | NVIDIA | $0.195/M | 133.6 tok/s | $0.086/M |
| #25 | Apertus 8B Instruct | Swiss AI Initiative | $0.200/M | n/a | $0.125/M |
| #26 | gpt-oss-20B (high) | OpenAI | $0.200/M | 240 tok/s | $0.088/M |
| #27 | gpt-oss-20B (low) | OpenAI | $0.200/M | 224.2 tok/s | $0.095/M |
| #28 | Ministral 3 14B | Mistral | $0.200/M | 106.9 tok/s | $0.200/M |
| #29 | NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning) | NVIDIA | $0.200/M | 87.3 tok/s | $0.088/M |
| #30 | Olmo 3 7B Instruct | Allen Institute for AI | $0.200/M | n/a | $0.125/M |
| #31 | Qwen2.5 Turbo | Alibaba | $0.200/M | 66.4 tok/s | $0.088/M |
| #32 | Qwen3 8B (Non-reasoning) | Alibaba | $0.200/M | 64.4 tok/s | $0.185/M |
| #33 | NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) | NVIDIA | $0.220/M | 133.6 tok/s | $0.096/M |
| #34 | Mistral 7B Instruct | Mistral | $0.225/M | 110.4 tok/s | $0.206/M |
| #35 | Mistral Small 3.1 | Mistral | $0.235/M | 158.2 tok/s | $0.138/M |
| #36 | Nova Lite | Amazon | $0.240/M | 191.7 tok/s | $0.105/M |
| #37 | Llama 3.2 Instruct 11B (Vision) | Meta | $0.245/M | 87.4 tok/s | $0.245/M |
| #38 | Gemma 3 27B Instruct | $0.250/M | n/a | $0.145/M |
| #39 | Granite 3.3 8B (Non-reasoning) | IBM | $0.250/M | 400.9 tok/s | $0.085/M |
| #40 | Granite 4.0 H Small | IBM | $0.250/M | 454.2 tok/s | $0.107/M |
| #41 | Llama 2 Chat 7B | Meta | $0.250/M | 100.6 tok/s | $0.100/M |
| #42 | Mistral Small 3.2 | Mistral | $0.250/M | 127.1 tok/s | $0.128/M |
| #43 | DeepSeek V4 Flash (Non-reasoning) | DeepSeek | $0.280/M | 97.4 tok/s | $0.175/M |
| #44 | DeepSeek V4 Flash (Reasoning, High Effort) | DeepSeek | $0.280/M | n/a | $0.175/M |
| #45 | DeepSeek V4 Flash (Reasoning, Max Effort) | DeepSeek | $0.280/M | 98.3 tok/s | $0.175/M |
| #46 | MiMo-V2.5 | Xiaomi | $0.280/M | 77.4 tok/s | $0.175/M |
| #47 | Gemma 3 12B Instruct | $0.290/M | n/a | $0.140/M |
| #48 | Qwen3 30B A3B (Non-reasoning) | Alibaba | $0.290/M | 68.6 tok/s | $0.133/M |
| #49 | Devstral Small (Jul '25) | Mistral | $0.300/M | 42.3 tok/s | $0.150/M |
| #50 | Gemma 4 12B (Reasoning) | $0.300/M | 158.6 tok/s | $0.150/M |
| #51 | Hermes 3 - Llama-3.1 70B | Nous Research | $0.300/M | 34.1 tok/s | $0.300/M |
| #52 | Ling 2.6 Flash | InclusionAI | $0.300/M | n/a | $0.150/M |
| #53 | MiMo-V2-Flash (Feb 2026) | Xiaomi | $0.300/M | 124.9 tok/s | $0.150/M |
| #54 | MiMo-V2-Flash (Non-reasoning) | Xiaomi | $0.300/M | 122.8 tok/s | $0.150/M |
| #55 | MiMo-V2-Flash (Reasoning) | Xiaomi | $0.300/M | 129.5 tok/s | $0.150/M |
| #56 | Nemotron 3 Nano Omni 30B A3B Reasoning | NVIDIA | $0.300/M | 276.7 tok/s | $0.131/M |
| #57 | Step 3.5 Flash | StepFun | $0.300/M | 217.5 tok/s | $0.150/M |
| #58 | Step 3.5 Flash 2603 | StepFun | $0.300/M | 231 tok/s | $0.150/M |
| #59 | Gemini 2.5 Flash-Lite (Non-reasoning) | $0.400/M | 229.5 tok/s | $0.175/M |
| #60 | Gemini 2.5 Flash-Lite (Reasoning) | $0.400/M | 265.2 tok/s | $0.175/M |
| #61 | Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning) | $0.400/M | n/a | $0.175/M |
| #62 | Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning) | $0.400/M | n/a | $0.175/M |
| #63 | Gemma 4 26B A4B (Non-reasoning) | $0.400/M | 48 tok/s | $0.198/M |
| #64 | Gemma 4 26B A4B (Reasoning) | $0.400/M | n/a | $0.198/M |
| #65 | Gemma 4 31B (Non-reasoning) | $0.400/M | 54.9 tok/s | $0.205/M |
| #66 | GLM-4.7-Flash (Non-reasoning) | Z AI | $0.400/M | 105.6 tok/s | $0.153/M |
| #67 | GLM-4.7-Flash (Reasoning) | Z AI | $0.400/M | 94.1 tok/s | $0.153/M |
| #68 | GPT-4.1 nano | OpenAI | $0.400/M | 118.2 tok/s | $0.175/M |
| #69 | GPT-5 nano (high) | OpenAI | $0.400/M | 150.4 tok/s | $0.138/M |
| #70 | GPT-5 nano (medium) | OpenAI | $0.400/M | 167 tok/s | $0.138/M |
| #71 | GPT-5 nano (minimal) | OpenAI | $0.400/M | 153.9 tok/s | $0.138/M |
| #72 | Hermes 4 - Llama-3.1 70B (Non-reasoning) | Nous Research | $0.400/M | 84.9 tok/s | $0.198/M |
| #73 | Hermes 4 - Llama-3.1 70B (Reasoning) | Nous Research | $0.400/M | 87.2 tok/s | $0.198/M |
| #74 | Jamba 1.5 Mini | AI21 Labs | $0.400/M | n/a | $0.250/M |
| #75 | Jamba 1.6 Mini | AI21 Labs | $0.400/M | 185.9 tok/s | $0.250/M |
| #76 | Llama Nemotron Super 49B v1.5 (Non-reasoning) | NVIDIA | $0.400/M | 43.9 tok/s | $0.175/M |
| #77 | Llama Nemotron Super 49B v1.5 (Reasoning) | NVIDIA | $0.400/M | 44.2 tok/s | $0.175/M |
| #78 | Qwen2.5 Instruct 72B | Alibaba | $0.400/M | n/a | $0.370/M |
| #79 | Qwen3 30B A3B 2507 Instruct | Alibaba | $0.400/M | 105.2 tok/s | $0.213/M |
| #80 | DeepSeek V3.2 Exp (Non-reasoning) | DeepSeek | $0.415/M | n/a | $0.310/M |