What the evidence says
- Developer
- DeepSeek
- Released
- April 24, 2026
- Domains
- Language
- Tasks
- Language modeling/generation · Question answering
- Access
- Open weights (unrestricted)
- Weights
- Open weights
Across the boards
Each row is one named authority's own current view. Nothing is combined.
- LMArena overall
- 1436.1 · rank #73 · 44,220 battles · published July 20, 2026 · board
- LMArena coding arena
- 1482.2 · rank #73 · board
- LMArena creative writing arena
- 1409.3 · rank #66 · board
- LMArena instruction following arena
- 1428.6 · rank #68 · board
- LMArena hard prompts arena
- 1459.1 · rank #69 · board
- LMArena multi-turn arena
- 1453.4 · rank #59 · board
- LMArena long queries arena
- 1450.0 · rank #63 · board
- LMArena math arena
- 1425.9 · rank #87 · board
- EQ-Bench 3
- 1287.2 · rank #22
Platform facts
OpenRouter's platform listing, retrieved July 22, 2026. Routed platform prices, not first-party provider prices.
- Input price
- $0.094 per 1M tokens
- Output price
- $0.19 per 1M tokens
- Context length
- 1,048,576 tokens
- Listed on OpenRouter
- April 24, 2026
Common questions
What is DeepSeek-V4-Flash?
DeepSeek-V4-Flash is a DeepSeek model released April 24, 2026, according to Epoch AI's reviewed catalog record. The record classifies it under language.
How does DeepSeek-V4-Flash rank on LMArena?
In the LMArena style-controlled text snapshot published July 20, 2026, DeepSeek-V4-Flash holds official rank 73 with a rating of 1436.1 across 44,220 battles.
What does DeepSeek-V4-Flash cost?
OpenRouter lists it at $0.094 per million input tokens and $0.19 per million output tokens, retrieved July 22, 2026. Those are the platform's routed prices, not first-party provider prices.