Solar Pro 2
Korea's first frontier LLM: 31B hybrid reasoning, above GPT-4.1 on benchmarks.
About
Solar Pro 2 is Upstage's July 2025 flagship: a 31B-parameter model with a hybrid architecture offering Chat Mode for fast responses and Reasoning Mode for structured multi-step thinking. It became Korea's first globally recognized frontier model by scoring 58 on the Artificial Analysis Intelligence Index — five points above GPT-4.1 and ahead of Mistral Small, Llama 4 Maverick, and DeepSeek V3 — while Upstage claims it was trained on a tenth of the chip budget of rivals. It carries a 65K context window, tool calling, and strong Korean fluency with cultural awareness for finance, legal, and medical domains.
The honest picture: it's a closed, API-only model — no open weights, so no self-hosting or fine-tuning. The 'tenth of the chips' efficiency claim is Upstage's own framing and hard to verify independently. Outside Korean-language and document-heavy enterprise niches its ecosystem is small, and Artificial Analysis scores, while credible, are one lens among many.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 46; Trending score 70/100 from researcher assessment.