LLM API Pricing Comparison
Every major LLM API price in one table. Chinese models in green, Western models in white. All prices per 1M tokens in USD.
Last updated: March 2026. Source: official API pricing pages.
Cheapest flagship input
$0.028/M
Qwen3.5-Flash
Free tier available
$0/M
GLM-4.7-Flash — no daily limits
vs GPT-5.2 ($1.75 input)
6-60x cheaper
Chinese flagship models
| Model | Provider | Input $/M | Output $/M | Cache $/M | Context | vs GPT-5.2 | vs Claude Sonnet |
|---|---|---|---|---|---|---|---|
| GLM-4.7-Flash | Zhipu AI | Free | Free | Free | 128K | ∞ cheaper | ∞ cheaper |
| Qwen3-Turbo | Alibaba Cloud / Qwen | $0.007 | $0.028 | — | 131K | 100% cheaper | 100% cheaper |
| Qwen3.5-Flash | Alibaba Cloud / Qwen | $0.028 | $0.275 | — | 1000K | 98% cheaper | 99% cheaper |
| Qwen3-Plus | Alibaba Cloud / Qwen | $0.07 | $0.28 | — | 131K | 96% cheaper | 98% cheaper |
| Qwen3.5-Plus | Alibaba Cloud / Qwen | $0.11 | $0.66 | — | 1000K | 94% cheaper | 96% cheaper |
| ERNIE 4.5 Turbo | Baidu | $0.11 | $0.44 | $0.028 | 128K | 94% cheaper | 96% cheaper |
| GPT-4o Mini | OpenAI | $0.15 | $0.6 | $0.075 | 128K | — | — |
| GPT-5 Mini | OpenAI | $0.25 | $2 | $0.025 | 200K | — | — |
| Qwen3 VL 235B | Alibaba Cloud / Qwen | $0.26 | $0.9 | — | 131K | 85% cheaper | 91% cheaper |
| DeepSeek V3.2 | DeepSeek | $0.28 | $0.42 | $0.028 | 128K | 84% cheaper | 91% cheaper |
| DeepSeek V3.2 (Thinking) | DeepSeek | $0.28 | $0.42 | $0.028 | 128K | 84% cheaper | 91% cheaper |
| ERNIE X1 | Baidu | $0.28 | $1.1 | — | — | 84% cheaper | 91% cheaper |
| MiniMax M2.5 | MiniMax | $0.3 | $1.2 | — | 205K | 83% cheaper | 90% cheaper |
| Gemini 2.5 Flash | $0.3 | $2.5 | $0.03 | 1000K | — | — | |
| Qwen3-Max | Alibaba Cloud / Qwen | $0.34 | $1.37 | — | 262K | 81% cheaper | 89% cheaper |
| DeepSeek V3.2 Speciale | DeepSeek | $0.4 | $1.2 | — | 164K | 77% cheaper | 87% cheaper |
| DeepSeek R1 0528 | DeepSeek | $0.45 | $2.15 | — | 131K | 74% cheaper | 85% cheaper |
| Kimi K2 Thinking | Moonshot AI | $0.47 | $2 | — | 256K | 73% cheaper | 84% cheaper |
| DeepSeek Prover V2 | DeepSeek | $0.5 | $2.18 | — | 164K | 71% cheaper | 83% cheaper |
| DeepSeek R1 | DeepSeek | $0.55 | $2.19 | $0.14 | 131K | 69% cheaper | 82% cheaper |
| ERNIE 4.5 | Baidu | $0.55 | $2.2 | — | 128K | 69% cheaper | 82% cheaper |
| Kimi K2.5 | Moonshot AI | $0.6 | $2.5 | $0.15 | 256K | 66% cheaper | 80% cheaper |
| Kimi K2 | Moonshot AI | $0.6 | $2.5 | $0.15 | 256K | 66% cheaper | 80% cheaper |
| GLM-4.7 | Zhipu AI | $0.6 | $2.2 | $0.11 | 128K | 66% cheaper | 80% cheaper |
| ERNIE 5.0 | Baidu | $0.83 | $3.31 | — | 128K | 53% cheaper | 72% cheaper |
| GLM-5 | Zhipu AI | $1 | $3.2 | $0.2 | 205K | 43% cheaper | 67% cheaper |
| Claude Haiku 4.5 | Anthropic | $1 | $5 | $0.1 | 200K | — | — |
| GLM-5-Code | Zhipu AI | $1.2 | $5 | $0.3 | — | 31% cheaper | 60% cheaper |
| GPT-5 | OpenAI | $1.25 | $10 | $0.125 | 128K | — | — |
| Gemini 2.5 Pro | $1.25 | $10 | $0.125 | 2000K | — | — | |
| GPT-5.2 | OpenAI | $1.75 | $14 | $0.175 | 200K | — | — |
| GPT-4o | OpenAI | $2.5 | $10 | $1.25 | 128K | — | — |
| Claude Sonnet 4.6 | Anthropic | $3 | $15 | $0.3 | 200K | — | — |
| Claude Opus 4.6 | Anthropic | $5 | $25 | $0.5 | 200K | — | — |
| GPT-5.2 Pro | OpenAI | $21 | $168 | $2.1 | 200K | — | — |
Methodology: All prices are from official API pricing pages as of March 2026. "vs GPT-5.2" and "vs Claude Sonnet" columns compare input token prices only. Actual cost depends on your input/output ratio and whether you use caching. Some providers offer tiered pricing based on request size — we show the base tier.
Need help choosing?
Check our model pages for detailed benchmarks, IDE setup guides, and access instructions.
Browse All Models →