DeepSeek v4 Flash 0423
DeepSeekBalanced
This model is no longer available on Elosia. This page is kept for informational purposes.
ThinkingTool UseStructured Output
About this model
DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...
Performance Tier
Balanced
DeepSeek v4 Flash 0423 is a balanced model from DeepSeek : strong performance at a reasonable price.
Strong cost-performance ratio. Reliable for most professional use cases without premium pricing.
Pricing
This model is included in Elosia plansEco
Minimal cost. Ideal for very high volume or simple tasks.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.140 |
| Output (completion) | $0.280 |
| Cache read | $0.028 |
Capabilities
Context Length1.0M
Max Output Tokens393K
TokenizerDeepSeek
Inputtext
Outputtext
Release DateApril 24, 2026
Benchmarks
General Intelligence
MMLU
88.7%
MMLU-Pro
86.2%
GPQA Diamond
88.1%
Mathematics
MATH-500
Not reported
Programming
HumanEval
69.5%
SWE-bench Verified
79%
LiveCodeBench
91.6%
Agentic
Terminal-Bench 2.0
56.9%
Where does DeepSeek v4 Flash 0423 stand?
Compare its performance index against every other model.
Recommended Use Cases
CodingMathematicsAnalysisGeneral Chat
Strengths
- Reasoning and code generation are the declared focus, with a reported suite weighted toward science questions, competitive programming and repository patches
- MoE 284B total / 13B active, a ratio that keeps throughput high for agents and coding assistants
- 1M-token context served by hybrid sparse attention
- Open-weight MIT, deployable on-prem
Limitations
- Limited factual knowledge density, which shows on recall-heavy tasks
- Max reasoning mode adds significant latency, so it does not suit real-time UX
- Creative writing is the weak axis of a model tuned for reasoning and code
Frequently asked questions
Resources
This model may use your data for training
