Nova 2 Lite (JP) vs DeepSeek V4 Pro
2026 Stored listed rates & illustrative token-cost comparison
📌 Executive Summary (TL;DR):Under a standard baseline of 1M input + 200K output tokens, listed rates for Nova 2 Lite (JP) and DeepSeek V4 Pro are virtually identical. Engineering teams should prioritize latency percentiles and contextual accuracy.(Price records last updated 2026-10-11 (UTC))
💡 10M Monthly Token Workload Simulation
Simulating a production workload (10M input + 2M output tokens): Nova 2 Lite (JP) listed rate totals ~$10.58, while DeepSeek V4 Pro totals ~$10.56. Switching to DeepSeek V4 Pro yields ~$0.02/mo in raw token savings.
Metric
Nova 2 Lite (JP) (AWS Bedrock)
DeepSeek V4 Pro (DeepSeek)
Input listed rate
$0.40 / 1M Tokens
$0.66 / 1M Tokens
Output listed rate
$3.31 / 1M Tokens
$1.98 / 1M Tokens
Cached Input rate
$0.10 / 1M Tokens
$0.02 / 1M Tokens
Context Window
1000K Tokens
1000K Tokens
Modality
llm
llm
🏆Related AI Scenario Leaderboards
Compare model standings across production scenarios weighted by editorial picks, public adoption, and verified price signals.
Want to map listed prices to your usage?
Enter token volume in the free estimator for a directional cost comparison. Validate quality, latency, and provider terms before migrating.