Log in
← All models

DeepSeek V3.2

Balanced

The proven open-weight value model — a staple of budget agent stacks.

DeepSeek
Maker
77/100
Overall
84
Value score
$0.26
Blended /1M
131K
Context
Open
Weights
Pricing: $0.23 in · $0.34 out / 1M tokens · synced from OpenRouter · updated 2026-06-22

Scorecard

Output Quality7.6/10 · 71% community
Agentic Ability7.4/10 · 70% community
Speed7.8/10 · 72% community
Value for $9.4/10 · 79% community
Reliability8.6/10 · 75% community

DeepSeek V3.2 is the well-loved previous-generation value model that remains a staple for agents that want dependable quality at very low cost. Open weights and a low price keep it popular even after V4's arrival.

Strengths
  • Excellent quality-per-dollar
  • Open weights
  • Battle-tested in agent stacks
Trade-offs
  • Shorter 128K context
  • Superseded by V4 on quality

Agents that run DeepSeek V3.2 well

Dependable open-weight output on a tight budget.

Best model for ClineBest model for AiderBest model for ContinueBest model for OpenClawBest model for OpenCode

DeepSeek V3.2 FAQ

Is DeepSeek V3.2 good for AI agents?
Dependable open-weight output on a tight budget. On our editorial scoring it rates 7.4/10 for agentic ability and 7.6/10 for output quality. Excellent quality-per-dollar.
How much does DeepSeek V3.2 cost?
Via OpenRouter, DeepSeek V3.2 is priced at $0.23 in · $0.34 out / 1M tokens — a blended rate of about $0.26 per 1M tokens for typical input-heavy agent use.
What is DeepSeek V3.2's context window?
DeepSeek V3.2 has a 131K-token context window. DeepSeek is the maker.
Is DeepSeek V3.2 open source?
Yes — DeepSeek V3.2 ships with open weights, so you can self-host it for privacy or run it through a hosted provider like OpenRouter. That openness is a big reason it's popular in open agent stacks.
What are the downsides of DeepSeek V3.2?
The main trade-offs: shorter 128k context; superseded by v4 on quality. It's strongest for dependable open-weight output on a tight budget.

DeepSeek V3.2 compared

DeepSeek V3.2 vs Sonnet 4.6DeepSeek V3.2 vs GPT-5.4DeepSeek V3.2 vs Gemini 3.5 FlashDeepSeek V3.2 vs GPT-5.3-CodexDeepSeek V3.2 vs Kimi K2.7 CodeDeepSeek V3.2 vs MiniMax M3DeepSeek V3.2 vs Qwen3.7 PlusDeepSeek V3.2 vs Devstral 2

The model is half the story — the agent is the other half

The model picks the moves; the agent runs the loop, the tools, and the guardrails. Once you've chosen a model, see which agent gets the most out of it.

Compare AI agents →