Claude Haiku 4.5 vs DeepSeek V4 Flash
Side-by-side API pricing and Artificial Analysis performance for Claude Haiku 4.5 (Anthropic) and DeepSeek V4 Flash (DeepSeek). Green cells mark the better value in each row.
Prices updated Jul 22, 2026, 3:44 AM UTC · refreshed hourly
| Claude Haiku 4.5 Anthropic | DeepSeek V4 Flash DeepSeek | |
|---|---|---|
| Pricing | ||
| Input $/1M | $1.00 | $0.094 |
| Output $/1M | $5.00 | $0.188 |
| Blended $/1M (3:1) | $2.00 | $0.117 |
| RAG example (30K in / 2K out) | $0.04 | $0.003189 |
| Quality & speed | ||
| Intelligence Index | 29.6 | 40.3 |
| Coding Index | 43.9 | 56.2 |
| Agentic Index | 16.4 | 31.1 |
| Output speed (tok/s) | — | 115 |
| Time to first token | — | 0.86s |
| Specs | ||
| Context window | 200K | 1.05M |
| Input modalities | text, image, file | text |
FAQ
Which is cheaper, Claude Haiku 4.5 or DeepSeek V4 Flash?
DeepSeek V4 Flash has the lower blended API price at $0.117 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 costs about $0.003189 per request versus $0.04. Also compare context windows: Claude Haiku 4.5 offers 200K and DeepSeek V4 Flash offers 1.05M.
Which scores higher on Artificial Analysis benchmarks?
DeepSeek V4 Flash leads on the Artificial Analysis Intelligence Index (40.3 vs 29.6). Check coding and agentic indices on this page for workload-specific tradeoffs.