DeepSeek

Large language model

DeepSeek V4 Flash 0731

Near-FrontierOpen WeightsReasoningMIT
Download Model →API Docs →
Input Price
$0.07
/M tokens
Output Price
$0.18
/M tokens
Intelligence
35
/100
Coding Rank
#24
Arena
Context
1.31072M
tokens
Parameters
284B MoE (13B active)

Verdict

One9Founders Verdictexcellent

July 31 post-train of V4-Flash. MIT 284B/13B MoE. Intelligence 35 on Index v4.3 — GLM-5.3 Flash now leads this band. Still a cheap MIT option.

“Still cheap. No longer the price-performance leader.”

Best For

  • ✓bootstrapped production inference
  • ✓high-volume MIT self-host
  • ✓price/performance

Not Great For

  • ✗enterprise certifications
  • ✗DPDP without self-hosting

Startup Recommendation

Best default for Indian bootstrapped backends if China-hosted API is acceptable, else self-host.

Pricing Details

Input
$0.07
per 1M tokens
Output
$0.18
per 1M tokens
Cached Input
$0.02
per 1M tokens

Capabilities

✓Function Calling
✓Structured Output
✗Vision
-Web Search
-Code Execution
-MCP
✗Audio Input
✗Video Input
✓Reasoning
✓Streaming
-Batch API
-Computer Use

Chatbot Arena Rankings

Coding
#24

Security & Compliance

✗SOC 2 Type II
✗GDPR
✗HIPAA
✗Enterprise SSO
✗Audit Logs

Training data policy: N/A — open weights. Self-host for full control.

DPDP Act (India): Self-host in India for full DPDP compliance.

Data residency: Self-hosted

India Availability

Self-host in India

Similar Models

Meta

Muse Spark 1.1

Arena #10$1.25 /M
Google

Gemini 3.7 Flash

Arena #11$0.75 /M
Google

Gemini 3.1 Pro Preview

Arena #15$2.00 /M
Anthropic

Claude Opus 4.8

Arena #18$5.00 /M
OpenAI

GPT-5.5

Arena #19$5.00 /M
Google

Gemini 3.6 Flash

Arena #21$0.75 /M