name = "DeepSeek V4 Pro"
description = "Flagship DeepSeek model for coding, reasoning, and agentic work"
family = "deepseek-thinking"
release_date = "2026-04-24"
last_updated = "2026-04-24"
attachment = false
reasoning = true
reasoning_options = [{ type = "effort", values = ["high", "max"] }]
temperature = true
knowledge = "2025-05"
tool_call = true
structured_output = true
open_weights = true

[interleaved]
field = "reasoning_content"

[cost]
input = 1.74
output = 3.48
cache_read = 0.0145

[limit]
context = 1_000_000
output = 384_000

[modalities]
input = ["text"]
output = ["text"]