High-volume, cost-sensitive work
Choose DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 costs 26% less on a typical mix of three input tokens for every output token ($0.15 against $0.20 per million tokens).
Choose Claude Haiku 5.5 for reused prompts and documents (RAG, agents), images as input and the latest release. Choose DeepSeek V4 Flash 0731 for high-volume, cost-sensitive work and writing and code generation.
Each answer below comes straight from the published prices, specifications and independent benchmark results.
Choose DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 costs 26% less on a typical mix of three input tokens for every output token ($0.15 against $0.20 per million tokens).
Choose DeepSeek V4 Flash 0731
DeepSeek V4 Flash 0731 charges $0.24 per million output tokens, against $0.50 for Claude Haiku 5.5. Output dominates the bill when answers are long.
Choose Claude Haiku 5.5
Claude Haiku 5.5 charges $0.01 per million cached input tokens, against $0.02 for DeepSeek V4 Flash 0731.
Choose Claude Haiku 5.5
Claude Haiku 5.5 accepts images as input; DeepSeek V4 Flash 0731 does not.
Choose Claude Haiku 5.5
Claude Haiku 5.5 is the newer release (October 2026, against Jul 31, 2026 for DeepSeek V4 Flash 0731).
| Volume | Claude Haiku 5.5 | DeepSeek V4 Flash 0731 | Difference |
|---|---|---|---|
| Prototype: 1M input, 250K output tokens | $0.23 | $0.18 | DeepSeek V4 Flash 0731, $0.05 less |
| Production: 20M input, 5M output tokens | $4.50 | $3.57 | DeepSeek V4 Flash 0731, $0.93 less |
| Scale: 500M input, 100M output tokens | $100.00 | $83.30 | DeepSeek V4 Flash 0731, $16.70 less |
Standard rates, without caching or batch discounts. Volumes are per month.
| Claude Haiku 5.5 | DeepSeek V4 Flash 0731 | |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Category | Language models | Language models |
| Status | Generally available | Generally available |
| Headline price | $0.20 / 1M tokens | $0.1488 / 1M tokens |
| Input | $0.10 / 1M tokens | $0.119 / 1M tokens |
| Output | $0.50 / 1M tokens | $0.238 / 1M tokens |
| Cached input | $0.01 / 1M tokens | $0.0238 / 1M tokens |
| Context window | 1M | 1.05M |
| Max output | 128k | 944k |
| Released | October 2026 | Jul 31, 2026 |
| Knowledge cutoff | June 2026 | No data |
| Licence | Proprietary | Proprietary |
| Capabilities | Vision, Reasoning, Tool calling, Structured output, Prompt caching, Batch API | Reasoning, Tool calling, Structured output, Prompt caching |
| API model id | claude-haiku-5-5 | deepseek-v4-flash-0731 |
Leader in bold.
| Benchmark | Claude Haiku 5.5 | DeepSeek V4 Flash 0731 |
|---|---|---|
| Intelligence Index | 43.4 | No data |
| SciCode | 55% | No data |
| Humanity's Last Exam | 44.4% | No data |
| Long-context reasoning (AA-LCR) | 82.7% | No data |
Independent evaluations by Artificial Analysis, run the same way for every model, at the reasoning setting shown. Percentages unless marked as an index.
Choose Claude Haiku 5.5 for reused prompts and documents (RAG, agents), images as input and the latest release. Choose DeepSeek V4 Flash 0731 for high-volume, cost-sensitive work and writing and code generation.
DeepSeek V4 Flash 0731 is 26% cheaper than Claude Haiku 5.5 at the headline price ($0.1488 / 1M tokens vs $0.20 / 1M tokens).
DeepSeek V4 Flash 0731 reads more per request: 1.05M tokens against 1M.
Claude Haiku 5.5 is made by Anthropic and DeepSeek V4 Flash 0731 by DeepSeek. DeepSeek V4 Flash 0731 is 26% cheaper than Claude Haiku 5.5 at the headline price ($0.1488 / 1M tokens vs $0.20 / 1M tokens). DeepSeek V4 Flash 0731 reads more per request: 1.05M tokens against 1M. Claude Haiku 5.5 is the newer model (October 2026, against Jul 31, 2026). Only Claude Haiku 5.5 offers vision, batch api.
At 20 million input and 5 million output tokens a month, Claude Haiku 5.5 costs $4.50 and DeepSeek V4 Flash 0731 costs $3.57, at standard rates.
Claude Haiku 5.5 was released on October 2026 and DeepSeek V4 Flash 0731 on Jul 31, 2026.