DeepSeek V4 Flash Latest vs Claude Haiku 4.5
Side-by-side API pricing and Artificial Analysis performance for DeepSeek V4 Flash Latest (~deepseek) and Claude Haiku 4.5 (Anthropic). Green cells mark the better value in each row.
Prices updated Sep 19, 2026, 9:54 AM UTC · refreshed hourly
| DeepSeek V4 Flash Latest ~deepseek | Claude Haiku 4.5 Anthropic | |
|---|---|---|
| Pricing | ||
| Input $/1M | $0.04 | $1.00 |
| Output $/1M | $0.08 | $5.00 |
| Blended $/1M (3:1) | $0.05 | $2.00 |
| RAG example (30K in / 2K out) | $0.00136 | $0.04 |
| Quality & speed | ||
| Intelligence Index | — | 17.6 |
| Coding Index | — | 43.9 |
| Agentic Index | — | 10.3 |
| Output speed (tok/s) | — | — |
| Time to first token | — | — |
| Specs | ||
| Context window | 1.31M | 200K |
| Input modalities | text | text, image, file |
FAQ
Which is cheaper, DeepSeek V4 Flash Latest or Claude Haiku 4.5?
DeepSeek V4 Flash Latest has the lower blended API price at $0.05 per 1M tokens (3:1 input:output mix), versus $2.00 for the other model. Prices are live from OpenRouter and refresh about hourly.
Which is better for RAG workloads?
For a typical RAG request (30K input / 2K output tokens), DeepSeek V4 Flash Latest costs about $0.00136 per request versus $0.04. Also compare context windows: DeepSeek V4 Flash Latest offers 1.31M and Claude Haiku 4.5 offers 200K.
Which scores higher on Artificial Analysis benchmarks?
Claude Haiku 4.5 leads on the Artificial Analysis Intelligence Index (17.6 vs —). Check coding and agentic indices on this page for workload-specific tradeoffs.