⌘K

DeepSeek V4 Pro

DeepSeek's new flagship reasoning model — frontier-class quality at open-weight pricing.

Official page ↗by DeepSeek

Intelligence

AA Intelligence Index44

independent benchmark score from Artificial Analysis, not Timace's own judgment

Best for

✓ Cost-sensitive frontier-class reasoning and coding

Specifications

Developer: DeepSeek
confidence 100%
www.deepseek.com
Verified Jul 21, 2026 · curated
Context window (tokens): 1M tokens
confidence 90%
artificialanalysis.ai
Verified Jul 21, 2026 · curated
Input $/MTok: $0.435 / MTok
confidence 85%
artificialanalysis.ai
Verified Jul 21, 2026 · curated
Output $/MTok: $0.87 / MTok
confidence 85%
artificialanalysis.ai
Verified Jul 21, 2026 · curated
Modalities: text, code
confidence 85%
www.deepseek.com
Verified Jul 21, 2026 · curated
API available: Yes
confidence 95%
www.deepseek.com
Verified Jul 21, 2026 · curated
Open source: Yes
confidence 80%
www.deepseek.com
Verified Jul 21, 2026 · curated

Compare against

Side-by-side →

Claude Fable 5Anthropic's most capable widely released model, for demanding reasoning and long-horizon agentic work.

Claude Haiku 4.5Near-frontier quality at commodity prices and high speed.

Claude Opus 4.8Anthropic's deepest reasoner for the hardest problems.

Claude Sonnet 5Anthropic's workhorse — frontier coding and agents at mid-tier pricing.