DeepSeek · deepseek-v4-flash
Before
$0.14
After
$0.22
Before capture
deepseek-v4-flash</td><td>deepseek-v4-pro</td></tr><tr><td colspan="2">BASE URL (OpenAI Format)</td><td colspan="2"><a href="https://api.deepseek.com" target="_blank" rel="noopener noreferrer">https://api.deepseek.com</a></td></tr><tr><td colspan="2">BASE URL (Anthropic Format)</td><td colspan="2"><a href="https://api.deepseek.com/anthropic" target="_blank" rel="noopener noreferrer">https://api.deepseek.com/anthropic</a></td></tr><tr><td colspan="2" style="text-align:center">MODEL VERSION</td><td>DeepSeek-V4-Flash-0731</td><td>DeepSeek-V4-Pro-0813</td></tr><tr><td colspan="2">THINKING MODE</td><td colspan="2">Supports both non-thinking and thinking (default) modes<br>See <a href="/guides/thinking_mode">Thinking Mode</a> for how to switch</td></tr><tr><td colspan="2">CONTEXT LENGTH</td><td colspan="2">1M</td></tr><tr><td colspan="2">MAX OUTPUT</td><td colspan="2">MAXIMUM: 384K</td></tr><tr><td rowspan="6">FEATURES</td><td><a href="/guides/json_mode">Json Output</a></td><td>✓</td><td>✓</td></tr><tr><td><a href="/guides/tool_calls">Tool Calls</a></td><td>✓</td><td>✓</td></tr><tr><td><a href="/guides/responses_api">Responses API</a></td><td>✓</td><td>✓</td></tr><tr><td><a href="/guides/anthropic_api">Anthropic API</a></td><td>✓</td><td>✓</td></tr><tr><td><a href="/guides/chat_prefix_completion">Chat Prefix Completion（Beta）</a></td><td>✓</td><td>✓</td></tr><tr><td><a href="/guides/fim_completion">FIM Completion（Beta）</a></td><td>Non-thinking mode only</td><td>Non-thinking mode only</td></tr><tr><td rowspan="3">PRICING<sup>(1)</sup></td><td>1M INPUT TOKENS (CACHE HIT)</td><td>$0.0028</td><td>$0.003625</td></tr><tr><td>1M INPUT TOKENS (CACHE MISS)</td><td>$0.14
- Captured
- 8/16/2026, 8:08:49 AM
- HTML hash
- 93eb81633c73b09aac937879fe128e3ff8cc173a3d6afa859116452bf223038a
After capture
deepseek-v4-flash</td><td>deepseek-v4-pro</td></tr><tr><td colspan="3">BASE URL (OpenAI Format)</td><td colspan="2"><a href="https://api.deepseek.com" target="_blank" rel="noopener noreferrer">https://api.deepseek.com</a></td></tr><tr><td colspan="3">BASE URL (Anthropic Format)</td><td colspan="2"><a href="https://api.deepseek.com/anthropic" target="_blank" rel="noopener noreferrer">https://api.deepseek.com/anthropic</a></td></tr><tr><td colspan="3" style="text-align:center">MODEL VERSION</td><td>DeepSeek-V4-Flash-0731</td><td>DeepSeek-V4-Pro-0813</td></tr><tr><td colspan="3">THINKING MODE</td><td colspan="2">Supports both non-thinking and thinking (default) modes<br>See <a href="/guides/thinking_mode">Thinking Mode</a> for how to switch</td></tr><tr><td colspan="3">CONTEXT LENGTH</td><td colspan="2">1M</td></tr><tr><td colspan="3">MAX OUTPUT</td><td colspan="2">MAXIMUM: 384K</td></tr><tr><td rowspan="6">FEATURES</td><td colspan="2"><a href="/guides/json_mode">Json Output</a></td><td>✓</td><td>✓</td></tr><tr><td colspan="2"><a href="/guides/tool_calls">Tool Calls</a></td><td>✓</td><td>✓</td></tr><tr><td colspan="2"><a href="/guides/responses_api">Responses API</a></td><td>✓</td><td>✓</td></tr><tr><td colspan="2"><a href="/guides/anthropic_api">Anthropic API</a></td><td>✓</td><td>✓</td></tr><tr><td colspan="2"><a href="/guides/chat_prefix_completion">Chat Prefix Completion（Beta）</a></td><td>✓</td><td>✓</td></tr><tr><td colspan="2"><a href="/guides/fim_completion">FIM Completion（Beta）</a></td><td>Non-thinking mode only</td><td>Non-thinking mode only</td></tr><tr><td rowspan="6">PRICING<sup>(1)</sup></td><td rowspan="2">1M INPUT TOKENS (CACHE HIT)</td><td>OFF-PEAK</td><td>$0.007</td><td>$0.022</td></tr><tr><td>PEAK</td><td>$0.014</td><td>$0.044</td></tr><tr><td rowspan="2">1M INPUT TOKENS (CACHE MISS)</td><td>OFF-PEAK</td><td>$0.22
- Captured
- 8/17/2026, 8:08:49 AM
- HTML hash
- b3b2eed30367fbca8f4b466b236b5e8f0f7cd761910e831efa5ca737f709eae0
Highlighting identifies the stored extracted fragments used by the parser. The hash is the fingerprint of the fetched source HTML for that monitor run.