[Integrations](https://spiral.ai/integrations) › [Devin](https://spiral.ai/integrations/devin) › [DeepSeek Pro Latest](https://spiral.ai/integrations/devin/deepseek-pro-latest)

# How to Use DeepSeek Pro Latest in Devin

Use DeepSeek Pro Latest for code changes, complex reasoning, and analysis across long texts. It currently points to DeepSeek V4 Pro 0813 and follows future Pro updates, so its underlying model can change.

|  |  |
| --- | --- |
| Model | DeepSeek Pro Latest |
| Available through | OpenRouter |
| Category | LLM |
| Price | Up to $0.042604 per run at the default inputs. $0.12987 per 1M input tokens, $10.40 per 1M output tokens. |
| Returns | text, json |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with DeepSeek Pro Latest in Devin

- Use DeepSeek Pro Latest with high reasoning effort to investigate the failing tests in this project. Trace the failure through the relevant files, make a focused fix, and run the affected tests.
- Use DeepSeek Pro Latest with max reasoning effort to review this project's retry logic for duplicate writes. Identify a concrete failure sequence, add a regression test, and propose a patch.
- Use DeepSeek Pro Latest with high reasoning effort to compare this project's implementation with its migration guide. List affected files, order the changes by dependency, and implement the first independently testable stage.
- Use DeepSeek Pro Latest to review the current diff against the project's requirements. Focus on incorrect behavior and missing edge cases; cite the relevant file and function for each finding.
- Use DeepSeek Pro Latest with low reasoning effort to summarize the test results as JSON. Follow this example: {"passed":12,"failed":1,"failures":[{"test":"rejects expired token","cause":"expiration check missing"}]}. Use actual results and validate the JSON.

## How to set up DeepSeek Pro Latest in Devin

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

### Use Spiral's MCP server

1. **Add a custom MCP**: Open Customize › MCPs. Click + and choose New MCP. Fill in:

2. **Sign in to Spiral and approve**: In Customize › MCPs, open Spiral and choose Connect. Spiral opens: sign in, pick the organization and click Authorize. No API key.

3. **Check it**: On Spiral's configuration page, click Test tools. Devin lists Spiral's tools.

### Use the Spiral skill

1. **Send Devin 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.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run DeepSeek Pro Latest**: One request; the result comes back by polling or a webhook.

```bash
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-pro-latest",
    "input": {
      "messages": [
        {
          "role": "system"
        }
      ]
    }
  }'
```

Then paste one of the prompts above: Devin finds DeepSeek Pro Latest on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/devin

## DeepSeek Pro Latest tools and prices

- [DeepSeek: DeepSeek Pro Latest](https://spiral.ai/tools/openrouter/~deepseek/deepseek-pro-latest) (OpenRouter): Up to $4.089452 per run at the default inputs. $0.12987 per 1M input tokens, $10.40 per 1M output tokens.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| seed |  | — |  |
| stop |  | — |  |
| min_p | 0 to 1 | — |  |
| tools |  | — |  |
| top_k | 0 to … | — |  |
| top_p | 0 to 1 | — |  |

## What can you do with DeepSeek Pro Latest in Devin?

DeepSeek Pro Latest can help trace bugs, plan migrations, and implement changes across a codebase. Give it the relevant files and acceptance criteria, then have the agent verify changes with tests.

### Fix a bug across several files

Trace a failing test through the relevant code, propose a focused patch, and verify the behavior with the project's tests.

### Plan a code migration

Compare existing code with supplied migration documentation. Identify affected calls, order the changes, and define checks for each stage.

### Compare long specifications

Find conflicting requirements across supplied specifications and build a decision table with references to the relevant sections.

### Investigate an incident

Analyze text logs and incident notes to reconstruct a timeline, rank possible causes, and identify evidence needed to test each explanation.

### Check a mathematical argument

Review a proposed solution for unsupported assumptions, test boundary cases, and explain corrections with a concise, checkable derivation.

### Extract structured records

Turn supplied support messages or operational notes into JSON using a worked example. Validate the records and flag missing information.

### Prepare a technical decision memo

Compare proposed designs against explicit requirements. Summarize tradeoffs, dependencies, and unanswered questions using only the supplied evidence.

### Review recurring operational changes

Compare successive text reports, separate new issues from repeated ones, and produce a short list of changes requiring attention.

## Overview

- [OpenRouter](https://spiral.ai/providers/openrouter)
- [All Devin guides](https://spiral.ai/integrations/devin)

## DeepSeek Pro Latest in other agents

- [Amp](https://spiral.ai/integrations/amp/deepseek-pro-latest)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/deepseek-pro-latest)
- [Claude](https://spiral.ai/integrations/claude-desktop/deepseek-pro-latest)
- [Claude Code](https://spiral.ai/integrations/claude-code/deepseek-pro-latest)
- [Cline](https://spiral.ai/integrations/cline/deepseek-pro-latest)
- [Codex](https://spiral.ai/integrations/codex/deepseek-pro-latest)
- [Cursor](https://spiral.ai/integrations/cursor/deepseek-pro-latest)
- [dots](https://spiral.ai/integrations/dots/deepseek-pro-latest)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/deepseek-pro-latest)
- [Gemini](https://spiral.ai/integrations/gemini/deepseek-pro-latest)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/deepseek-pro-latest)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/deepseek-pro-latest)
- [goose](https://spiral.ai/integrations/goose/deepseek-pro-latest)
- [Grok](https://spiral.ai/integrations/grok/deepseek-pro-latest)
- [Grok Build](https://spiral.ai/integrations/grok-build/deepseek-pro-latest)
- [Hermes](https://spiral.ai/integrations/hermes/deepseek-pro-latest)
- [Instinct](https://spiral.ai/integrations/instinct/deepseek-pro-latest)
- [Manus](https://spiral.ai/integrations/manus/deepseek-pro-latest)
- [Muse](https://spiral.ai/integrations/muse/deepseek-pro-latest)
- [Muse Code](https://spiral.ai/integrations/muse-code/deepseek-pro-latest)
- [n8n](https://spiral.ai/integrations/n8n/deepseek-pro-latest)
- [Notion](https://spiral.ai/integrations/notion/deepseek-pro-latest)
- [OpenClaw](https://spiral.ai/integrations/openclaw/deepseek-pro-latest)
- [opencode](https://spiral.ai/integrations/opencode/deepseek-pro-latest)
- [Perplexity](https://spiral.ai/integrations/perplexity/deepseek-pro-latest)
- [Raycast](https://spiral.ai/integrations/raycast/deepseek-pro-latest)
- [VS Code](https://spiral.ai/integrations/vscode/deepseek-pro-latest)
- [Windsurf](https://spiral.ai/integrations/windsurf/deepseek-pro-latest)
- [Zapier](https://spiral.ai/integrations/zapier/deepseek-pro-latest)

## More Devin guides

- [DeepSeek Flash Latest](https://spiral.ai/integrations/devin/deepseek-flash-latest)
- [DeepSeek R1 0528](https://spiral.ai/integrations/devin/deepseek-r1-0528)
- [DaisyRose 27B](https://spiral.ai/integrations/devin/daisyrose-27b)
- [DeepSeek V3](https://spiral.ai/integrations/devin/deepseek-v3)
- [Cydonia 24B V4.1](https://spiral.ai/integrations/devin/cydonia-24b-v4-1)
- [DeepSeek V3 0324](https://spiral.ai/integrations/devin/deepseek-v3-0324)

## Questions about DeepSeek Pro Latest in Devin

**Can Devin use DeepSeek Pro Latest?**

Yes. Add Spiral to Devin as an MCP server, and Devin can find and run DeepSeek Pro Latest like any built-in tool.

**How much does DeepSeek Pro Latest cost in Devin?**

Up to $0.042604 per run at the default inputs. $0.12987 per 1M input tokens, $10.40 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Devin plan is separate.

**Do I need an OpenRouter account to use DeepSeek Pro Latest in Devin?**

No. Spiral runs DeepSeek Pro Latest 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 Pro Latest in Devin?**

About two minutes: add a custom MCP, then sign in to Spiral and approve, then check it.

**Does Devin wait while DeepSeek Pro Latest runs?**

DeepSeek Pro Latest runs in the background. Devin starts the run, checks on it and gives you the text, json when it's ready.

**What is DeepSeek Pro Latest good at?**

It is a useful choice for code changes, complex reasoning, and analysis across long texts. DeepSeek's published evaluations cover these areas; check the result against your tests or source material.

**Does DeepSeek Pro Latest always use the same model?**

No. It currently points to DeepSeek V4 Pro 0813, but the name follows future Pro updates. Recheck representative tasks after an update if consistent behavior matters.

**How much text can DeepSeek Pro Latest handle?**

The current target supports a 1,048,576-token context window and up to 393,216 completion tokens. Those limits belong to DeepSeek V4 Pro 0813 and can change when the target changes.

**Can DeepSeek Pro Latest understand screenshots or charts?**

The current Pro target does not support vision. Supply extracted text or a table of values for analysis.

**Which reasoning setting should I choose?**

DeepSeek recommends low effort for simple tasks, high for daily agent work, and max for more complex problems. Start with high for debugging or planning, then compare max on the hardest cases.

