Spiral gives your agent every tool with one account. Browse the tools →

opencode · LLMReleased 2025-12-01

How to Use DeepSeek V3.2 in opencode

DeepSeek V3.2 suits coding, math, and research that combines reasoning with tools. Unlike DeepSeek V3.2 Speciale, it supports tool calls in both thinking and non-thinking modes.

Prompts to try with DeepSeek V3.2 in opencode

Copy an example into opencode, or ask in your own words.

Use DeepSeek V3.2 with reasoning enabled to investigate the failing tests in this project. Reproduce one failure, trace its cause, make the smallest fix, and run the relevant tests.

DeepSeek V3.2opencode
Try asking
  • Use DeepSeek V3.2 to review the algorithm in the current diff. Check correctness, time complexity, and empty or oversized inputs; give a concrete counterexample for each suspected bug.

  • Use DeepSeek V3.2 with reasoning enabled to compare the cache implementations in this repository. Trace invalidation behavior and recommend a change only if you can demonstrate a correctness problem.

  • Use DeepSeek V3.2 to turn the bug report in this project's issue notes into a minimal reproduction. Add a regression test, propose a focused patch, and report what you verified.

  • Use DeepSeek V3.2 with reasoning disabled to draft release notes from the current diff. Group changes by user impact and include only behavior supported by the code.

How to set up DeepSeek V3.2 in opencode

Pick the way that suits you. Each takes about two minutes.

Use Spiral's MCP server

  1. Add the server

    Leave out --global to add Spiral to the current project only. No header and no client ID or secret: opencode registers itself with Spiral.

    opencode mcp add spiral --global --url https://api.spiral.ai/mcp
  2. Sign in

    Your browser opens. Sign in to Spiral, pick the organization and click Authorize, then return to the terminal. No API key. Inside opencode, /mcps starts the same sign-in.

    opencode mcp auth spiral
  3. Check it

    Spiral shows as connected.

    opencode mcp list

Use the Spiral skill

  1. Send opencode this message

    Your agent reads Spiral's skill and sets itself up.

    set up https://spiral.ai/SKILL.md
  2. Sign in or paste a key

    Approve Spiral in the browser, or give the agent a key from the console.

Use the Spiral CLI

  1. Install and sign in

    Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

    npx -y spiral-cli@latest init

Call the Spiral API

  1. Create an API key

    In the console, under API keys.

  2. Run DeepSeek V3.2

    One request; the result comes back by polling or a webhook.

    curl https://api.spiral.ai/v1/runs \
      -H "Authorization: Bearer $SPIRAL_API_KEY" \
      -H "Content-Type: application/json" \
      -H "Idempotency-Key: $(uuidgen)" \
      -d '{
        "provider": "openrouter",
        "model": "deepseek/deepseek-v3.2",
        "input": {
          "messages": [
            {
              "role": "system"
            }
          ]
        }
      }'

Then paste one of the prompts above: opencode finds DeepSeek V3.2 on Spiral, shows the price and runs it.

DeepSeek V3.2 providers and prices

Spiral runs DeepSeek V3.2 through OpenRouter. Open a card for today's price, with every fee included.

OpenRouterLLM
DeepSeek: DeepSeek V3.2
deepseek/deepseek-v3.2

Settings you can ask for

Ask opencode for any of these in your prompt.

SettingAllowed valuesDefaultDescription
Seeda whole number
Min p0 to 1
Top ka whole number from 0
Top p0 to 1
Logprobson or off
Max tokens1 to 147456

The most output tokens to generate.

Temperature0 to 2
Top logprobs0 to 20
Presence penalty-2 to 2
Frequency penalty-2 to 2
Repetition penalty0 to 2

Every API field of DeepSeek: DeepSeek V3.2 is on its tool page: open its card above.

What can you do with DeepSeek V3.2 in opencode?

DeepSeek V3.2 can investigate failures, review algorithms, and reason through changes across project files. Give it a reproducible problem and require a patch backed by relevant checks.

Investigate a failing test

Use DeepSeek V3.2 to connect a failure trace to the relevant code, propose a small fix, and check it against the original failure.

Review an algorithm

Ask DeepSeek V3.2 to examine correctness, runtime, and edge cases before proposing an alternative implementation.

Check a math solution

Have DeepSeek V3.2 identify the first unsupported step in a proposed solution and verify the result with a separate calculation.

Research a technical question

Use DeepSeek V3.2 with search tools to gather primary sources, compare their claims, and produce an answer with links and unresolved questions.

Compare technical proposals

Give DeepSeek V3.2 the proposals and explicit requirements. Request a comparison that links each recommendation to supporting text and flags missing evidence.

Triage recurring incidents

Use DeepSeek V3.2 to examine new incident descriptions, identify possible duplicates, and suggest the next diagnostic check for each unresolved issue.

Check calculations with code

Pair DeepSeek V3.2 with a code execution tool to test formulas, inspect discrepancies, and explain which assumptions affect the result.

Questions about DeepSeek V3.2 in opencode

Can opencode use DeepSeek V3.2?

Yes. Add Spiral to opencode as an MCP server, and opencode can find and run DeepSeek V3.2 like any built-in tool.

How much does DeepSeek V3.2 cost in opencode?

Up to $0.002036 per run at the default inputs. $0.338 per 1M input tokens, $0.494 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your opencode plan is separate.

Do I need an OpenRouter account to use DeepSeek V3.2 in opencode?

No. Spiral runs DeepSeek V3.2 on its own OpenRouter account. You sign in to Spiral once and never paste a key.

How long does it take to set up DeepSeek V3.2 in opencode?

About two minutes: add the server, then sign in, then check it.

Does opencode wait while DeepSeek V3.2 runs?

DeepSeek V3.2 runs in the background. opencode starts the run, checks on it and gives you the text, json when it's ready.

What is DeepSeek V3.2 useful for?

Coding and math are practical starting points: DeepSeek evaluates V3.2 on both. It also supports tasks that combine reasoning with tool calls, such as investigating a bug or checking sources.

When should I enable thinking?

Try thinking for problems with several dependencies, such as debugging or comparing conflicting evidence. Try non-thinking for straightforward extraction, then compare accuracy on your own examples; tool calls work in either mode.

How much context can DeepSeek V3.2 handle?

The published model configuration specifies 163,840 token positions. That figure describes context capacity, not a promise that a single answer can contain that many tokens.

What are DeepSeek V3.2's main weaknesses?

DeepSeek acknowledges gaps in world knowledge compared with leading proprietary models and identifies token efficiency as a limitation. Supply source material for factual tasks and request a concise final answer.

How does DeepSeek V3.2 differ from DeepSeek V3.2 Speciale?

DeepSeek V3.2 supports reasoning alongside tool calls. DeepSeek V3.2 Speciale is designed for deep reasoning and does not support tool calling.

Get started

Start using DeepSeek V3.2 in opencode

One Spiral account for every tool your agent needs. No subscription; you pay per run.

Paste this into opencode:

Set up opencode

set up https://spiral.ai/SKILL.md
Are you an AI agent? See setup options