Upstage

About Upstage

Upstage builds powerful large language models and document processing engines to transform workflows and empower leading businesses like yours.

Built and refined by our team of AI experts, our technology brings the highest standards of accuracy and performance to your business.

Convert PDFs, scans, and emails into clean, machine-readable text ready for AI pipelines.

Pull structured key-value data from invoices, claims, and contracts with audited accuracy.

Website ↗Documentation ↗

Source ↗ · 2026-11-12T02:17:08.162722083Z

16 indexed models

16 matching models

Upstage: Solar Pro 4

upstage/solar-pro4-20260810

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 work, and coding. When using BYOK with ZDR enforcement, th

Context window
524,288
Input modalities
text
Output modalities
text
Capabilities
promptCaching, reasoning, structuredOutputs, toolCalling

Benchmarks

  • Artificial Analysis: 28.2
  • LMArena: 1,388.197
  • OpenRouter Benchmarks: 80.808%

Inference prices

  • Published price : $0.03 / cache_read_tokens_per_million · openRouter ↗
  • Published price : $0.15 / input_tokens_per_million · openRouter ↗
  • Published price : $0.60 / output_tokens_per_million · openRouter ↗

Upstage: Solar Pro 3

upstage/solar-pro-3

Solar Pro 3 is Upstage's powerful Mixture-of-Experts (MoE) language model. With 102B total parameters and 12B active parameters per forward pass, it delivers exceptional performance while maintaining computational efficiency. Optimized...

Context window
131,072
Input modalities
text
Output modalities
text

Benchmarks

  • Artificial Analysis: 7.8

upstage/SOLAR-10.7B-Instruct-v1.0

vertexAI/publishers/hf-upstage/models/solar-10.7b-instruct-v1.0@001

Input modalities
text
Output modalities
text

Benchmarks

  • LMArena: 1,083.093
  • Open LLM Leaderboard v1: 74.201%
  • Open LLM Leaderboard v2: 20.572%

upstage/Llama-2-70b-instruct

vertexAI/publishers/hf-upstage/models/llama-2-70b-instruct@001

Input modalities
text
Output modalities
text

Benchmarks

  • Open LLM Leaderboard v1: 67.375%

upstage/llama-30b-instruct

vertexAI/publishers/hf-upstage/models/llama-30b-instruct@001

Input modalities
text
Output modalities
text

Benchmarks

  • Open LLM Leaderboard v1: 58.911%

upstage/llama-30b-instruct-2048

vertexAI/publishers/hf-upstage/models/llama-30b-instruct-2048@001

Input modalities
text
Output modalities
text

Benchmarks

  • Open LLM Leaderboard v1: 60.908%

upstage/llama-65b-instruct

vertexAI/publishers/hf-upstage/models/llama-65b-instruct@001

Input modalities
text
Output modalities
text

Benchmarks

  • Open LLM Leaderboard v1: 64.507%

upstage/SOLAR-0-70b-16bit

vertexAI/publishers/hf-upstage/models/solar-0-70b-16bit@001

Input modalities
text
Output modalities
text

Benchmarks

  • Open LLM Leaderboard v1: 70.106%

upstage/SOLAR-10.7B-v1.0

vertexAI/publishers/hf-upstage/models/solar-10.7b-v1.0@001

Input modalities
text
Output modalities
text

Benchmarks

  • Open LLM Leaderboard v1: 66.038%
  • Open LLM Leaderboard v2: 16.855%

Upstage: Solar Decide

upstage/solar-decide-20260928

Solar Decide is Upstage's structured decision model, served as a System One endpoint on Solar Mini 4. Send a state along with typed questions, and it returns a choice, a score, or a yes/no answer, each with a calibrated probability taken directly from the model rather than writte

Context window
524,288
Input modalities
text
Output modalities
decisions
Capabilities
decisions, promptCaching

Inference prices

  • Published price : $0.05 / cache_read_tokens_per_million · openRouter ↗
  • Published price : $0.05 / input_tokens_per_million · openRouter ↗

Upstage: Solar Decide Flash

upstage/solar-decide-flash-20261008

Solar Decide Flash is Upstage's low-latency structured decision model, a faster variant of [Solar Decide](/upstage/solar-decide) built on Solar Mini 4 and served through the System One (`/v1/systemone`) API. Instead of generating text, it reads a `state` and answers typed questio

Context window
524,288
Input modalities
text
Output modalities
decisions
Capabilities
decisions, promptCaching

Inference prices

  • Published price : $0.05 / cache_read_tokens_per_million · openRouter ↗
  • Published price : $0.05 / input_tokens_per_million · openRouter ↗

Upstage: Solar Mini 4

upstage/solar-mini4-20260922

Solar Mini 4 is Upstage's compact, cost-efficient language model, a 35B-parameter mixture-of-experts with 3B active parameters and a 524K context window. It is built for agentic use cases where response speed and cost matter, with fluent Korean alongside strong English and Japane

Context window
524,288
Input modalities
text
Output modalities
text
Capabilities
promptCaching, reasoning, structuredOutputs, toolCalling

Benchmarks

  • Artificial Analysis: 24.1

Inference prices

  • Published price : $0.01 / cache_read_tokens_per_million · openRouter ↗
  • Published price : $0.05 / input_tokens_per_million · openRouter ↗
  • Published price : $0.20 / output_tokens_per_million · openRouter ↗

upstage/Solar-Open-100B

vertexAI/publishers/hf-upstage/models/solar-open-100b@001

Input modalities
text
Output modalities
text

upstage/TinySolar-248m-4k

vertexAI/publishers/hf-upstage/models/tinysolar-248m-4k@001

Input modalities
text
Output modalities
text

upstage/TinySolar-248m-4k-code-instruct

azureFoundry/HuggingFace/upstage-tinysolar-248m-4k-code-instruct@1

`upstage/TinySolar-248m-4k-code-instruct` powered by Text Generation Inference (TGI) - [Example Notebook](https://aka.ms/hugging-face-text-generation-streaming-online-endpoint) - [Original Model Card](https://huggingface.co/upstage/TinySolar-248m-4k-code-instruct) - [Text Genera

Input modalities
text
Output modalities
text

Models

Loading
EVIDENCE / COMPARISON

Benchmarks

Explore every benchmark

Loading

These charts contain model-linked results. The charts above also include unlinked source profiles. Compare scores within each evaluation. Model quality, usage, transparency and system performance measure different things.

Install TensorTrace