Skip to main content

Best value: Spicy Index for the money

Higher means a better Spicy Index, further left is cheaper. Highlighted models are unbeaten on both at once.

Video models: Spicy Index vs price

Spicy Index against price for Video models
0255075100$0.10$0.20$0.50$1.00$2.00Price per reference generation (log scale)Spicy IndexWan 3.0LTX-2 19BWan 2.6 Flash

Prices use list rates for a 5-second reference clip, at the available resolution nearest 720p with audio off where supported. Models billed per clip show that clip rate. The selected tier appears beside each price.

Best value picks

No other model both scores higher and costs less than these. Pick along this line by budget.

  1. LTX-2 19B57.8 · $0.096 per 5-second clip
  2. Wan 2.6 Flash78.1 · $0.113 per 5-second clip
  3. Wan 3.087.7 · $0.45 per 5-second clip

Image models: Spicy Index vs price

Spicy Index against price for Image models
0255075100$0.002$0.005$0.01$0.02$0.05Price per image (log scale)Spicy IndexQwen Image 2.1 LoRAQwen Image 2.1 ProQwen Image 2.1 TurboGPT Image 2.5 SunburstP-Image LoRAZ-Image Spicy

Prices use one image at the available tier nearest 1K, taking the lowest list rate within that tier. Open a model for the quote for your settings.

Best value picks

No other model both scores higher and costs less than these. Pick along this line by budget.

  1. GPT Image 2.5 Sunburst62.1 · $0.00182 per image
  2. P-Image LoRA71.3 · $0.006 per image
  3. Z-Image Spicy73.7 · $0.0124 per image
  4. Qwen Image 2.1 Turbo81.1 · $0.0144 per image
  5. Qwen Image 2.1 LoRA81.9 · $0.03 per image
  6. Qwen Image 2.1 Pro82.0 · $0.036 per image

Language models: Spicy Index vs price

Spicy Index against price for Language models
0255075100$0.20$1.00$5.00$20.00Price per 1M mixed tokens (3 input : 1 output) (log scale)Spicy IndexDeepSeek V4.1 FlashClaude Haiku 5.5Claude Opus 5.5

The comparison price is 750,000 input tokens plus 250,000 output tokens at the base rates (3:1). Actual input and output usage is billed at their respective model rates.

Best value picks

No other model both scores higher and costs less than these. Pick along this line by budget.

  1. Claude Haiku 5.576.5 · $0.20 per 1M mixed tokens (3 input : 1 output)
  2. DeepSeek V4.1 Flash81.9 · $0.525 per 1M mixed tokens (3 input : 1 output)
  3. Claude Opus 5.582.5 · $7.20 per 1M mixed tokens (3 input : 1 output)