OpenRating is an independent, open-source AI rating platform. Join the community

OOpenRating
All models

Upstage · Released Feb 5, 2026

Solar Open2 250B

Upstage's 250B parameter open-weights model featuring Depth-Up Scaling (DUS) and strong multilingual reasoning.

Open weights

OpenRating Index

37.4

0–100 composite

Reasoning Index

54.0

0–100 composite

Cost / Task

$0.19

benchmark run

Output speed

94 tok/s

blended tok/s

Time to first token

0.6s

median latency

Input price

$0.14

per 1M tokens

Output price

$0.56

per 1M tokens

Context window

128K

tokens

Strengths

  • Depth-Up Scaling architecture for parameter efficiency

  • Bilingual Korean/English and multilingual performance

  • Open weights for custom fine-tuning pipelines

How we rate Solar Open2 250B

OpenRating scores every model on the same axes. The OpenRating Index is a weighted composite of reasoning, knowledge, and agentic benchmarks, normalised to a 0–100 scale. Speed is blended output measured in tokens per second under standardised load, with time-to-first-token reported separately. Prices are list input and output prices per million tokens, and Task Cost measures end-to-end workload spending.

OpenRating Index

37.4

Reasoning Index

54.0

Token Pricing

Input per 1M tokens

$0.14

Output per 1M tokens

$0.56

Cost of 1M input + 1M output

$0.7

Task Cost Economics

Empirical cost breakdown for a standard OpenRating Index benchmark task ($0.19 total).

Reasoning (Chain of Thought)
$0.038
Answer generation
$0.018
Cache write overhead
$0.072
Cache hit queries
$0.052
Input payload
$0.005