DeepSeek V4 Flash
LLM Provider / Model ยท Ultra-cheap high-quality coding, bulk classification, context-heavy tasks
At a glance
| Input price | $0.14 |
| Output price | $0.28 |
| Cache price | $0.003 |
| Price tier | Budget |
| Context window | 1M+ |
| Max output | 384K |
| Context tier | 500K+ |
| Speed tier | Fast |
| Latency | medium |
| Knowledge cutoff | Jun 2025 |
| Modality | Text-only |
| Model ID | deepseek-v4-flash |
| Provider | DeepSeek |
| HumanEval | 92% |
| MMLU | 86% |
| SWE-Bench | 48% |
| Benchmark | high (coding) |
| Released | 2026 |
| Hosting | Closed/API |
| Capabilities | Tool use, Streaming, Structured output, Thinking mode, Prefix completion, FIM |
What DeepSeek V4 Flash does
Tool use, Streaming, Structured output, Thinking mode, Prefix completion, FIM
Best for
Ultra-cheap high-quality coding, bulk classification, context-heavy tasks
Works well with
Integration
Agent / Orchestration
Conflicts & caveats
- Trae's built-in model menu may not include DeepSeek V4 Flash โ Trae primarily ships with Claude/GPT/Gemini/DeepSeek.
- Google Antigravity is built around Gemini 3 Pro; Claude and recent GPT-5 models also work, but DeepSeek V4 Flash is not in the supported model list.
- Kiro's spec mode is tuned for Claude models โ DeepSeek V4 Flash may work but spec-driven flows are best with Claude.
- OpenAI Codex (VS Code) is tied to OpenAI's GPT-5.1 Codex family โ DeepSeek V4 Flash is not selectable inside this integration.
- Gemini CLI is locked to Google Gemini models โ your LLM is DeepSeek V4 Flash.
- Claude Code only works with Claude models โ your LLM is DeepSeek V4 Flash.
- Devin uses Cognition's proprietary model internally โ your selected DeepSeek V4 Flash will be ignored.
- OpenAI Codex (Cloud) runs GPT-5.1 Codex internally โ your selected DeepSeek V4 Flash will be ignored.
- Jules runs Gemini models internally โ your selected DeepSeek V4 Flash will be ignored.
- The Claude Code SDK is tuned for Claude models โ DeepSeek V4 Flash may work via the model adapter but loses Claude-specific features (thinking, prompt caching, sub-agents).
- Lovable uses a fixed internal model โ your selected DeepSeek V4 Flash will be ignored.
- Replit Agent 3 uses its own model selection internally โ your selected DeepSeek V4 Flash will be ignored.
Build a full stack around DeepSeek V4 Flash โ Flowpicker shows compatibility warnings before you commit.
Open the stack planner โ