Solar LLM
Upstage's Solar LLMs — Korea's frontier family, now Solar Pro 3
Overview
Upstage's Solar family runs from the Apache-2.0 Solar 10.7B (2023) to Solar Pro 4 (April 2026), a 128K-context MoE served via API at $0.30/$1.20 per million tokens. It offers the strongest Korean-language performance of any commercial model family.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 2 Sep 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- +10% bonus credits
- Community support
- +15% bonus credits
- Email support (3-day)
- +20% bonus credits
- Priority support (1-day)
- All models incl. Document AI
- Private/on-prem deployment
- Dedicated 24/7 support
Use cases
What you can produce with Solar LLM
- A production-grade Korean/English/Japanese chatbot API integrated into your enterprise stack via OpenAI-compatible endpoint
- Automated document intelligence pipeline that parses, classifies, and extracts structured data from PDFs at $0.01–$0.03 per page
- Fine-tuned Solar model variant for a specific domain (finance, legal, healthcare) hosted on-premise for data-residency compliance
- Agentic workflow system using Solar Pro 3's tool-use and structured-output capabilities for multi-step business process automation
- Benchmark comparison report validating Solar Pro 2 performance vs GPT-4o and DeepSeek R1 on your organisation's specific task distribution
- Korean NLP evaluation suite benchmarked against Ko-MMLU, Hae-Rae, and Ko-Arena-Hard-Auto datasets
Solar is the rare 2023 open-weights story that grew into a real product line. The Apache-2.0 Solar 10.7B and its depth up-scaling trick topped the Hugging Face leaderboard in December 2023; the commercial line then kept shipping — Solar Pro 2 (31B, July 2025) scored above GPT-4.1 on Artificial Analysis's Intelligence Index, the first Korean-built model rated frontier-class, and Solar Pro 3 (January 2026) is a 128K-context mixture-of-experts model at $0.15/$0.60 per million tokens with a free window until March 2026. Korean-language performance is the standout, and Upstage's document-AI products give it real APAC enterprise traction. Caveats: a smaller ecosystem than the US majors, and the old 10.7B open weights are far behind current open models — the API line is where the capability lives.
What people say
Fifteen cents per million input tokens, 60 for output — that is Solar Pro 3 pricing at its January 2026 launch, and Upstage gave it away free until March 2 to pull developers in.
Depth up-scaling — splicing extra transformer layers into a pretrained base and continuing training — put Solar 10.7B briefly ahead of Mistral 7B and Llama 2 13B, and those Apache-2.0 weights are still on Hugging Face. But Upstage kept shipping: Solar Pro at 22B in late 2024, then Solar Pro 2 at 31B in July 2025, which Artificial Analysis scored at 58 on its Intelligence Index — above GPT-4.1, an unusually large result for its size class and the first Korean-built model it rated frontier-class. Solar Pro 3 moved to a mixture-of-experts design, 102B total with 12B active, a 128K context window, selectable chat and reasoning modes and, per Upstage, roughly double the agentic task performance of its predecessor.
On Ko-MMLU, Hae-Rae and Arena-Hard-style benchmarks the Pro models match or beat the US giants, and Upstage pairs the LLM with a document-parse and OCR line doing real enterprise business in Korea and Japan. The complaints cluster on two things: the developer console is still maturing next to OpenAI's or Anthropic's, and language coverage is Korean, English and Japanese first — there is no native support for Bahasa, Thai, Vietnamese or Tagalog, aside from a one-off sovereign LLM project for Thailand's Jasmine Group.
Upstage is also headed toward Korea's first generative-AI IPO, with KB Securities and Mirae Asset running a process targeting a KOSPI listing as early as the second half of 2026, on revenue that reportedly grew about 320% over three years. Commercial tiers run on monthly usage commitments rather than flat SaaS plans, from Explore at $100+ to Scale at $5,000+, with a separate Enterprise track. For Southeast Asian teams the case is thinner — this is a Korea-first, APAC-second stack — but as a cost-efficient GPT-4-class API from outside the US-China duopoly it is worth benchmarking. Self-hosting the old 10.7B in 2026 is nostalgia; the capability lives in the API line.
Summary of public user & expert reviews, compiled by RECATOOLS.
Notable facts
- Solar 10.7B was trained by a South Korean company and outperformed all European and American models of similar size, demonstrating APAC's growing contributions to frontier AI.
- The Depth Upscaling technique used for Solar allows models to be built more efficiently by combining and extending existing pre-trained architectures.
- Solar was the first non-Western model to achieve #1 on the HuggingFace Open LLM Leaderboard.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including Solar LLM's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Solar LLM unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Solar LLM directly →
Spotted something out of date? Suggest an update →
More in LLMs & Chat
