Back to Models
Upstage
solar-pro4

Solar Pro 4

Solar Pro 4 is Upstage's cost-efficient large language model, featuring a 524K context window. It is built for long-horizon tasks and agentic workflows, with strong capabilities in office productivity, document-intensive...

8/10/2026
524,288 tokens
#102 Code (Overall)
Input: $0.099/M
Output: $0.396/M

Specifications

Modalities

Input
text
Output
text

Supported Parameters

frequency_penalty
include_reasoning
max_tokens
presence_penalty
reasoning
response_format
structured_outputs
temperature
tool_choice
tools
top_p

Max Output Tokens

131,072

Data policy

Prompt retention
May be retained
Training
Not used for training

Whether a provider trains on prompts is a declared value from our terms with them. Retention is determined from the upstream listing for every host this model can reach. Neither is guessed.

Measured performance

Time to first response (p50)449ms
Throughput (p50)90 tokens/s

    Measured through FastMetal's own gateway over the last 30 days (712 first-response samples).

    Throughput is sampled once a day: the median of 26 observations over the last 30 days.

    Time to first response is measured to the first streamed chunk; throughput counts all output tokens, including reasoning tokens. These are observations, not a performance guarantee.

    Code Examples

    curl https://api.fastmetal.ai/v1/chat/completions \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer YOUR_API_KEY" \
      -d '{
        "model": "solar-pro4",
        "messages": [{"role": "user", "content": "Hello!"}]
      }'

    How Solar Pro 4 actually answers

    Real responses to our standard prompts, recorded on FastMetal.

    Count the number of 'r's in 'strawberry'

    Count the number of 'r's in 'strawberry'. Explain your reasoning step by step.

    The word "strawberry" contains **3** 'r's. Here's the step-by-step breakdown: 1. **Spell out the word explicitly** to isolate each letter: `s - t - r - a - w - b - e - r - r - y` 2. **Scan through the letters in order** and count every instance of the lowercase letter `r`: - `s` → 0 - `t` → 0 - `r` → **1** - `a` → 0 - `w` → 0 - `b` → 0 - `e` → 0 - `r` → **2** - `r` → **3** - `y` → 0 3. **Total the count**: The three `r`s appear at positions 3, 8, and 9 in the 10-letter word. **Final count: 3**
    11.4s response215 output tokens

    Debug This Error

    I'm getting the following error in my Node.js application: TypeError: Cannot read properties of undefined (reading 'map') at UserList (/app/components/UserList.js:12:25) at renderWithHooks (/app/node_modules/rea…

    ### What's Causing the Error The error `TypeError: Cannot read properties of undefined (reading 'map')` means that the `users` prop being passed to your `UserList` component is `undefined`. Since you're calling `.map()` directly on it, JavaScript throws an error because `undefined` doesn't have a `map` method. This typically happens when: - The parent component hasn't passed the `users` prop - The prop is still loading (e.g., from an API call) - The prop is conditionally undefined --- ### How to Fix It You have a few options to handle this safely: #### ✅ Option 1: Add a Default Prop Value (Recommended) Provide a fallback empty array so `.map()` always has something to work with: ```j…
    24.7s response539 output tokens

    Code Review

    Please review the following Python function and suggest improvements for readability, performance, and best practices: def get_data(url, retries=3): import requests import time for i in range(retries):…

    Here’s a breakdown of the issues in the original function, followed by a refactored version and a detailed explanation of the improvements. ### 🔍 Key Issues in the Original Code | Area | Problem | |------|---------| | **Import organization** | `import` inside the function is evaluated on every call. Imports should live at the top of the module. | | **Exception handling** | Bare `except:` catches `SystemExit`, `KeyboardInterrupt`, and other fatal errors. Always catch specific exceptions (`requests.RequestException`). | | **Loose equality** | `== True` is redundant and can cause subtle bugs. Use `if item["active"]:` or `.get()`. | | **HTTP status handling** | Calling `.json()` only on `200`…
    17.5s response1172 output tokens

    Compare these answers side by side with other models →

    Frequently asked questions

    How much does the Solar Pro 4 API cost?
    On FastMetal, Solar Pro 4 is billed per token: $0.099 per 1M input tokens and $0.396 per 1M output tokens on a US-dollar account, before tax. Usage is drawn from a prepaid balance; there is no subscription or monthly fee.
    Can I call Solar Pro 4 with the OpenAI SDK?
    Yes. Point base_url at https://api.fastmetal.ai/v1 and pass "solar-pro4" as the model; existing OpenAI-style code works unchanged, including streaming, tool calls and structured output.
    What is the context window of Solar Pro 4?
    524,288 tokens, shared between the prompt and the response.
    How does Solar Pro 4 rank?
    #158 on the public arena's Japanese board (ELO 1,312). Ranks move as the leaderboard is updated.
    What do I need to try Solar Pro 4?
    Create an account and add credit; the model is then available both in the browser chat and over the API. There is no contract or minimum spend.

    Try Solar Pro 4 right now

    Solar Pro 4 is available on FastMetal through one API key. Start in the browser, or call it from the OpenAI SDK.

    Leaderboard

    Text
    russianELO: 1,358
    #194
    CodingELO: 1,461
    #136
    MathELO: 1,414
    #137
    Creative WritingELO: 1,311
    #219
    Instruction FollowingELO: 1,380
    #176
    Hard PromptsELO: 1,409
    #176
    Multi-TurnELO: 1,386
    #183
    OverallELO: 1,386
    #182
    JapaneseELO: 1,312
    #158
    ChineseELO: 1,415
    #171
    KoreanELO: 1,371
    #110
    EnglishELO: 1,419
    #151
    frenchELO: 1,397
    #166
    germanELO: 1,356
    #170
    spanishELO: 1,389
    #159
    Code
    OverallELO: 1,371
    #102

    Rankings: LMArena leaderboard dataset (lmarena-ai/leaderboard-dataset), CC BY 4.0, snapshot 2026-10-08; category names mapped and ratings rounded by FastMetal.