DeepSeek V4 Flash 0423 is 22.0× cheaper per million tokens (blended). DeepSeek V4 Flash 0423 has the larger context window (1,048,575 vs 200,000 tokens, 5.2×). DeepSeek V4 Flash 0423 is open-weights; the other is not.

Claude Haiku 4.5 vs DeepSeek V4 Flash 0423

Live catalog fields. Best-in-row is highlighted.

Claude Haiku 4.5

Anthropic · anthropic__claude-haiku-4.5

DeepSeek V4 Flash 0423open

DeepSeek · deepseek__deepseek-v4-flash

Identity

Fieldanthropic__claude-haiku-4.5deepseek__deepseek-v4-flash
ProviderAnthropicDeepSeek
Sluganthropic__claude-haiku-4.5deepseek__deepseek-v4-flash
Statuslivelive
Open weightsNoYes
License——
HuggingFace—deepseek-ai/DeepSeek-V4-Flash

Quality

Fieldanthropic__claude-haiku-4.5deepseek__deepseek-v4-flash
Quality index (0–100)——

Cost

Fieldanthropic__claude-haiku-4.5deepseek__deepseek-v4-flash
Input $ / 1M tokens$1.00$0.09
Cached input $ / 1M$0.10$0.02
Output $ / 1M tokens$5.00$0.18

Context

Fieldanthropic__claude-haiku-4.5deepseek__deepseek-v4-flash
Max context tokens200,0001,048,575
Max output tokens64,000943,717

Modalities

Fieldanthropic__claude-haiku-4.5deepseek__deepseek-v4-flash
Inputtext, image, filetext
Outputtexttext