TTPI$0.326−4.72%TPPI$2.23+18.13%TCPI$2.94+5.96%Nemotron 3.5 Lightningcached$0.025$0.023−4.30%Qwen3.5 397B A17Bcached$0.220$0.225+2.30%Nemotron 3 Superout$0.400$0.450+12.50%Z.ai: GLM 4.6Vcached$0.050$0.055+10.00%Thinking Machines: Inklingcached$0.160$0.170+6.30%Muse Glimmer 30Bout$1.50$1.20−20.00%MoonshotAI: Kimi K3cached$0.300$0.270−10.00%MoonshotAI: Kimi K2.6cached$0.094$0.073−22.50%MoonshotAI: Kimi K2.7 Codeout$3.30$3.35+1.50%DeepSeek V3.1 Terminusin$0.270$0.300+11.10%SpaceXAI: Grok 4.7cached$0.400$0.500+25.00%Qwen3 30B A3B Instruct 2507in$0.090$0.100+11.10%

TTPI$0.326−4.72%TPPI$2.23+18.13%TCPI$2.94+5.96%Nemotron 3.5 Lightningcached$0.025$0.023−4.30%Qwen3.5 397B A17Bcached$0.220$0.225+2.30%Nemotron 3 Superout$0.400$0.450+12.50%Z.ai: GLM 4.6Vcached$0.050$0.055+10.00%Thinking Machines: Inklingcached$0.160$0.170+6.30%Muse Glimmer 30Bout$1.50$1.20−20.00%MoonshotAI: Kimi K3cached$0.300$0.270−10.00%MoonshotAI: Kimi K2.6cached$0.094$0.073−22.50%MoonshotAI: Kimi K2.7 Codeout$3.30$3.35+1.50%DeepSeek V3.1 Terminusin$0.270$0.300+11.10%SpaceXAI: Grok 4.7cached$0.400$0.500+25.00%Qwen3 30B A3B Instruct 2507in$0.090$0.100+11.10%

BETA

← Back to prices·All Upstage models

UpstageEfficientLIVE PRICING

Solar Pro 3

text->text

Upstage's flagship Solar. 22B-class dense model; punches above its weight on Korean and English benchmarks.

Solar Pro 3 is a efficient AI model from Upstage. It costs $0.150 per million input tokens and $0.600 per million output tokens (blended $0.285/M), with a 131,072-token context window.

Profile inherited from upstream Solar Pro ↗ — this is a hosted variant of the same open-weights model.

Modalities textOfficial model page ↗Provider pricing ↗API docs ↗Pricing source ↗Compare with another model →Estimate monthly cost →

INPUT

$0.150/M

per million input tokens

OUTPUT

$0.600/M

per million output tokens

BLENDED 70/30

$0.285/M

unchanged since 3 May

default reference rate · how it is calculated →

CONTEXT

131,072

tokens

What it is good at

Strong Korean coverage
Single-GPU friendly
Open weights variants

Typical use cases

Korean-first production chat
Multilingual self-hosted chat

Benchmarks

vs. best public score

Scores inherited from Solar Pro — this is a hosted variant of the same open-weights model, so the underlying benchmark scores are identical.

Multitask academic knowledge across 57 subjects.

Python function synthesis from docstrings.

LMArena Elo1192 Elo

Crowd-sourced head-to-head preference Elo rating.

Hand-curated from each provider's published reports and public leaderboards. Methodology varies across sources, treat as directional rather than authoritative.

More from Upstage

Frontier · 32,000 ctx

in $1.500/Mout $6.000/M

Efficient · 32,000 ctx

in $0.150/Mout $0.600/M

Solar Embedding

Embedding · 4,000 ctx

in $0.100/M

Frequently asked questions

How much does Solar Pro 3 cost?

Solar Pro 3 costs $0.150 per million input tokens and $0.600 per million output tokens, for a blended reference rate of $0.285 per million tokens.

What is Solar Pro 3's context window?

Solar Pro 3 supports up to 131,072 tokens of context in a single request.

What is Solar Pro 3 best for?

Solar Pro 3 is well suited to Strong Korean coverage, Single-GPU friendly and Open weights variants.

Who makes Solar Pro 3?

Solar Pro 3 is developed and served by Upstage.

Terms used on this page

Full glossary →

Token Input tokens Output tokens Price per million tokens Context window Prompt caching Batch API Blended cost