V4-Flash
#244 in Frontier LLMsUnknown · v4 · 2× · last seen Aug 04, 2026
3
Momentum
DeepSeek-V4-Flash is a Mixture-of-Experts language model from DeepSeek with 284 billion total parameters and 13 billion activated parameters per Token, optimized for coding, tool usage, and agentic workflows. The model supports a 1-million-Token context window and is released as Open Source under the MIT license. The official version DeepSeek-V4-Flash-0731 replaced the Preview version introduced in April 2026 on July 31, 2026, and received significantly improved agent capabilities through new Post-Training, including native support for the Responses API and Codex integration.
Momentum trend
24.06.22.09.
Features
| Key Benchmark (%) | Terminal-Bench 2.1: 82.7 (official, vendor-reported) |
| Context Window (Tokens) | 1,048,576 tokens (1M), max output 384,000 tokens |
| License | MIT License (open weights) |
| Multimodality | Text-based model with function calling and reasoning mode; no official image/audio multimodality confirmed |
| Platform | DeepSeek API (deepseek-v4-flash), chat.deepseek.com, Hugging Face, LM Studio |
| Price per 1M Tokens | $0.14 input (cache miss) / $0.0028 input (cache hit) / $0.28 output |
| Release Date | Preview: April 24, 2026; official release (0731): July 31, 2026 |