Track how AI model API pricing changes over time. Data collected weekly since 2026-02-16, covering 331+ models across 51 providers.
Average input and output cost per 1M tokens across each provider's paid models.
| Provider | Models | Avg Input $/M | Avg Output $/M |
|---|---|---|---|
| Liquid AI | 3 | $0.017 | $0.053 |
| IBM | 1 | $0.017 | $0.110 |
| essentialai | 1 | $0.150 | $0.150 |
| StepFun | 1 | $0.100 | $0.300 |
| Allen AI | 7 | $0.139 | $0.343 |
| Microsoft | 2 | $0.343 | $0.380 |
| Meta | 13 | $0.432 | $0.560 |
| Tencent | 1 | $0.140 | $0.570 |
| Upstage | 1 | $0.150 | $0.600 |
| Baidu | 5 | $0.196 | $0.694 |
| NVIDIA | 7 | $0.327 | $0.694 |
| Inception | 3 | $0.250 | $0.750 |
| Meituan | 1 | $0.200 | $0.800 |
| Kuaishou | 1 | $0.207 | $0.828 |
| ByteDance | 5 | $0.155 | $0.980 |
| DeepSeek | 11 | $0.359 | $0.993 |
| Alibaba | 48 | $0.236 | $1.08 |
| arcee-ai | 5 | $0.475 | $1.13 |
| eleutherai | 1 | $0.800 | $1.20 |
| deepcogito | 1 | $1.25 | $1.25 |
Models that have decreased in price since tracking began.
| Model | Provider | Previous Output $/M | Current Output $/M | Change |
|---|---|---|---|---|
| Mistral Small 3.1 24B | Mistral AI | $0.560 | $0.110 | -80.4% |
| Qwen3 Next 80B A3B Thinking | Alibaba | $1.20 | $0.780 | -35.0% |
| Qwen3 Max | Alibaba | $6.00 | $3.90 | -35.0% |
| Qwen-Plus | Alibaba | $1.20 | $0.780 | -35.0% |
| Qwen3.5-Flash | Alibaba | $0.400 | $0.260 | -35.0% |
| Qwen VL Max | Alibaba | $3.20 | $2.08 | -35.0% |
| gpt-oss-20b | OpenAI | $0.140 | $0.110 | -21.4% |
| DeepSeek V3.2 | DeepSeek | $0.400 | $0.380 | -5.0% |
| MiniMax M2.5 | MiniMax | $1.20 | $1.17 | -2.5% |
| DeepSeek V3.1 Terminus | DeepSeek | $0.790 | $0.780 | -1.3% |
| Model | Provider | Previous Output $/M | Current Output $/M | Change |
|---|---|---|---|---|
| Command R (08-2024) | Cohere | $0.600 | $10.00 | +1566.7% |
| Qwen2.5 Coder 32B Instruct | Alibaba | $0.200 | $1.00 | +400.0% |
| Qwen3 235B A22B Thinking 2507 | Alibaba | $0.600 | $1.50 | +149.2% |
| QwQ 32B | Alibaba | $0.400 | $0.580 | +45.0% |
| Qwen3 30B A3B Thinking 2507 | Alibaba | $0.340 | $0.400 | +17.6% |
| Mistral Small 3.2 24B | Mistral AI | $0.180 | $0.200 | +11.1% |
| Gemma 3 27B | $0.150 | $0.160 | +6.7% | |
| Phi 4 | Microsoft | $0.140 | $0.140 | +0.0% |
| DeepSeek V3.2 | DeepSeek | $0.400 | $0.380 | +-5.0% |
AI model API pricing has been on a consistent downward trajectory since 2023. OpenAI's GPT-4 launched at $60/M output tokens; today, models with comparable capability cost under $5/M. This represents a 90%+ price reduction in under three years.
Key pricing trends observed across the industry:
Yes. AI API pricing has dropped dramatically since 2023. GPT-4-class performance that cost $60/M tokens in 2023 now costs under $3/M tokens. Competition from DeepSeek, Gemini, and open-source models continues to push prices lower.
Major providers typically adjust pricing 2-4 times per year, usually when launching new models. Price drops of 50-90% are common when a new generation replaces an older model.
Google and DeepSeek have been the most aggressive on pricing. Google offers free API access to Gemini Flash models, and DeepSeek offers frontier-quality reasoning at a fraction of competitor prices.