Side-by-side · API pricing
DeepSeek V4.1 Flash vs Claude Sonnet 5.5
DeepSeek V4.1 Flash's stored input rate is 100% lower than Claude Sonnet 5.5's ($0.003 vs $2.00 per 1M Tokens; matched scope: 1M Tokens · global · llm). DeepSeek V4.1 Flash's stored output rate is 76% lower than Claude Sonnet 5.5's ($2.40 vs $10.00 per 1M Tokens; matched scope: 1M Tokens · global · llm).
Price records last updated 2026-10-11 (UTC)
DeepSeek · Anthropic · Only like-for-like stored records are compared; verify current terms at a linked source where available.
Model A
DeepSeek V4.1 Flash
DeepSeek
- Input / 1M Tokens
- $0.003
- Output / 1M Tokens
- $2.40
- Region
- global
- Modality
- llm
- Source
- View stored source
- Context window
- 1,048,576
- Stored price records
- 1
Model B
Claude Sonnet 5.5
Anthropic
- Input / 1M Tokens
- $2.00
- Output / 1M Tokens
- $10.00
- Region
- global
- Modality
- llm
- Source
- View stored source
- Context window
- 1,000,000
- Stored price records
- 2
Per-platform pricing
Each row retains its own currency, unit, region, and source status. Rates with different scopes are not ranked against each other.
DeepSeek V4.1 Flash
- openrouterInput / 1M Tokens: $0.003 · Output / 1M Tokens: $2.40USD · 1M Tokens · global · llmSource
Frequently asked questions
- Which is cheaper, DeepSeek V4.1 Flash or Claude Sonnet 5.5?
- DeepSeek V4.1 Flash's stored input rate is 100% lower than Claude Sonnet 5.5's ($0.003 vs $2.00 per 1M Tokens; matched scope: 1M Tokens · global · llm). DeepSeek V4.1 Flash's stored output rate is 76% lower than Claude Sonnet 5.5's ($2.40 vs $10.00 per 1M Tokens; matched scope: 1M Tokens · global · llm).
- Where does this comparison data come from?
- This page uses stored price records. A provider or platform source is linked where the record contains one; records without a source link are explicitly marked. Record timestamps show database updates, not independent verification.
- What does this comparison include?
- It compares a price field only when both records have the same currency, billing unit, region, and modality. It does not normalize taxes, volume discounts, availability, latency, quality, or provider-specific terms.
Need a decision review for a real workload? Apply for the Founder Pilot.