رفتن به محتوای اصلی

جدول مدل‌های زبانی، قیمت و مزایای هریک

این جدول از ارزان‌ترین مدل به گرانترین مدل‌ها مرتب شده است

Model Developer Input Price ($/1M tokens) Cached Input Price ($/1M tokens) Output Price ($/1M tokens) Notes
GPT-4.1-nano OpenAI $0.10 $0.03 $0.40 (OpenAI) Optimized for cost and speed.
DeepSeek coder DeepSeek $0.14 N/A $0.28 (Artificial Analysis)  

QWEN 3

235b-a22b

Alibaba $0.14 N/A $0.60
Gemini 2.5 flash Google $0.15 $0.04 $0.60 (non-thinking) / $3.50 (thinking) (Google AI for Developers)
DeepSeek V3 DeepSeek $0.27 (miss) / $0.07 (hit) $0.07 $1.10 (DeepSeek API Docs) Good for conversational tasks. **Input price depends on cache hit/miss. ***Uses "Cache Hit" price.
Grok-3-mini xAI $0.30 N/A $0.50 (TechCrunch) Cost-effective option from xAI. (Beta)
GPT-4.1-mini OpenAI $0.40 $0.10 $1.60 (OpenAI) Mid-tier option.
DeepSeek R1 DeepSeek $0.55 (miss) / $0.14 (hit) $0.14 $2.19 (DeepSeek API Docs) Specialized reasoning model. **Input price depends on cache hit/miss. ***Uses "Cache Hit" price. Slower response time due to thinking.
GPT-4o-mini OpenAI $0.60 $0.30 $2.40 (OpenAI) Excellent balance of cost and performance for many tasks.
O3-mini OpenAI $1.10 $0.55 $4.40 (Artificial Analysis)
O4-mini OpenAI $1.10 $0.28 $4.40 (OpenAI)
Gemini 2.5 Pro Google $1.25 / $2.50* N/A $10.00 / $15.00* Google's top model. *Price depends on prompt length (>200k tokens is higher). Output cost includes "thinking" tokens. Context Caching planned. (Preview)
GPT-4.1 OpenAI $2.00 $0.50 $8.00 (OpenAI) High-performance text model.
Sonar Deep Research Perplexity AI $2.00 N/A $8.00 (reasoning tokens $3.00 - Price per 1000 Search Queries $5) (Perplexity) Specialized for deep research tasks. Price calculation complex (includes search costs). Caching not specified. (Using Legacy Pricing for consistency)
O1-mini OpenAI $3.00 N/A $12.00 (Ibbaka)  
Claude 3.7 Sonnet  Anthropic $3.00 $0.30 (90 % off) $15.00 (thinking tokens included) (Anthropic) Top-tier model from Anthropic. Price includes "thinking" tokens. †Uses "Prompt Caching - Cache Read" price.
Claude 3.7 Sonnet thinking Anthropic $3.00 $0.30 (90 % off) $15.00 (thinking tokens included) (Anthropic)  
Grok-3 xAI $3.00 N/A $15.00 Flagship model from xAI. (Beta)
GPT-4o OpenAI $2.50 $1.25 $10.00 OpenAI's flagship multimodal model.
O3 OpenAI $10.00 $2.50 $40.00 (OpenAI)  
gpt-image-1
N/A N/A N/A (image outputs billed per-image - $0.10 in average) (OpenAI)

Key for Notes/Symbols

Note: Cached input pricing mechanisms and availability vary by provider. Some offer direct discounts on cached tokens (OpenAI, Google), while others have specific cache hit/miss pricing (DeepSeek) or prompt caching features with read/write costs (Anthropic). Where a specific cached input price isn't offered or found, "N/A" is used.

  • * Price depends on prompt context length.

  • ** Input price depends on whether the input prefix is found in the cache (hit) or not (miss).

  • *** This is the "Cache Hit" price provided by DeepSeek, representing the cost for the cached portion of the input.

  • † Anthropic uses a "Prompt Caching" feature with separate "Cache Write" (higher cost) and "Cache Read" (lower cost) prices. The "Cache Read" price is listed here as the closest equivalent to discounted cached input.

This table provides a snapshot to help you choose the right tool for the job. Need something quick and affordable for drafting emails? GPT-4o-mini might be perfect. Tackling a complex data analysis or coding problem? O1, Claude 3.7 Sonnet, or Gemini 2.5 Pro could be better choices.


مقایسه مدل‌های جدید

🔍 راهنمای انتخاب مدل

برای کارهای پیچیده علمی و تحقیقاتی: O3 یا Gemini 2.5 Pro
برای استفاده روزمره با کیفیت بالا و هزینه کمتر: O4 mini یا Llama 4 Maverick
برای پاسخ‌های سریع و کارآمد: O4 mini یا Gemini 2.5 Flash
برای بهترین نسبت قیمت به کارایی: Llama 4 Maverick (قدرت O4 با قیمت O4 mini)


منبع بنچمارک AI‌ها: lmarena.ai

معرفی پلتفرم‌ها و LLM های مختلف: huggingface.co