DeepSeek V3 is 5.0× cheaper per million tokens (blended). Claude Haiku 4.5 has the larger context window (200,000 vs 163,840 tokens). DeepSeek V3 is open-weights; the other is not.
Claude Haiku 4.5 vs DeepSeek V3
Live catalog fields. Best-in-row is highlighted.
Claude Haiku 4.5
Anthropic · anthropic__claude-haiku-4.5
DeepSeek V3open
DeepSeek · deepseek__deepseek-chat
Identity
| Field | anthropic__claude-haiku-4.5 | deepseek__deepseek-chat |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Slug | anthropic__claude-haiku-4.5 | deepseek__deepseek-chat |
| Status | live | live |
| Open weights | No | Yes |
| License | — | — |
| HuggingFace | — | deepseek-ai/DeepSeek-V3 |
Quality
| Field | anthropic__claude-haiku-4.5 | deepseek__deepseek-chat |
|---|---|---|
| Quality index (0–100) | — | — |
Cost
| Field | anthropic__claude-haiku-4.5 | deepseek__deepseek-chat |
|---|---|---|
| Input $ / 1M tokens | $1.00 | $0.32 |
| Cached input $ / 1M | $0.10 | — |
| Output $ / 1M tokens | $5.00 | $0.89 |
Context
| Field | anthropic__claude-haiku-4.5 | deepseek__deepseek-chat |
|---|---|---|
| Max context tokens | 200,000 | 163,840 |
| Max output tokens | 64,000 | 16,384 |
Modalities
| Field | anthropic__claude-haiku-4.5 | deepseek__deepseek-chat |
|---|---|---|
| Input | text, image, file | text |
| Output | text | text |