Language modelActive

Solar Pro 4

Upstage

Released: August 11, 2026
Data date: September 8, 2026
Predecessor: Solar Pro 3

Speed: 83.2 tokens/s
Output throughput
Price: $0.3 · $1.2
Input · Output / 1M tokens
Input: Text
Output: Text

Solar Pro 4 is designed for verifiable work products rather than short chat answers. Upstage names documents, terminal tasks, web research, and multi-step tool use as core cases. It offers a 512K-token context, up to 128K output tokens, and English, Korean, and Japanese support; Solar Pro 3 is the previous generation.

Reasoning is enabled by default, its effort can be adjusted, and the reasoning trace can be returned. Upstage reports the largest gains on agentic tasks and work deliverables, but does not publish a parameter count or complete architecture. Solar Pro 4 is offered through Upstage Console and OpenRouter. It is a proprietary hosted service, not local weights.

Specifications and access

Specification	Value and source
Weights	ProprietarySource
Access	APISource
Context window	512,000 tokensSource
Notes	Agentic model with a 512K context window and up to 128K output tokens; parameter count is not published.Source
Release	2026-08-11Source
Architecture	not disclosedSource
Input	TextSource
Output	TextSource
Maximum output	128,000 tokensSource
Languages	English, Korean, JapaneseSource
Reasoning mode	YesSource

Measurements without matching peer values

These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.

AA-LCR (Show measurement, test conditions, and source)

Source value: 71
Score: 71
Metric: percent
Unit: %
Category: long-context
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page

Source official model card/release page

AIME 2026 (Show measurement, test conditions, and source)

Source value: 95.3
Score: 95.3
Metric: percent
Unit: %
Category: reasoning
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk row in-house

Source official model card/release page

Arena-Hard v2 (Japanese) (Show measurement, test conditions, and source)

Source value: 81.9
Score: 81.9
Metric: percent
Unit: %
Category: multilingual
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk row in-house

Source official model card/release page

BrowseComp (Show measurement, test conditions, and source)

Source value: 49.2
Score: 49.2
Metric: percent
Unit: %
Category: agentic
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk rows in-house

Source official model card/release page

GDPval-AA v2 (Show measurement, test conditions, and source)

Source value: 38.8
Score: 38.8
Metric: percent
Unit: %
Category: agentic
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk rows in-house

Source official model card/release page

GPQA Diamond (Show measurement, test conditions, and source)

Source value: 89
Score: 89
Metric: percent
Unit: %
Category: reasoning
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page

Source official model card/release page

KMMLU-Pro (Show measurement, test conditions, and source)

Source value: 79.2
Score: 79.2
Metric: percent
Unit: %
Category: multilingual
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk row in-house

Source official model card/release page

Ko-GDPval (Show measurement, test conditions, and source)

Source value: 87.3
Score: 87.3
Metric: percent
Unit: %
Category: multilingual
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk row in-house

Source official model card/release page

KorMedMCQA (Show measurement, test conditions, and source)

Source value: 93.2
Score: 93.2
Metric: percent
Unit: %
Category: multilingual
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk row in-house

Source official model card/release page

LiveCodeBench (Show measurement, test conditions, and source)

Source value: 87.8
Score: 87.8
Metric: percent
Unit: %
Category: coding
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk row in-house

Source official model card/release page

MCP-Atlas (Show measurement, test conditions, and source)

Source value: 61.4
Score: 61.4
Metric: percent
Unit: %
Category: agentic
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk rows in-house

Source official model card/release page

MMLU-Pro (Show measurement, test conditions, and source)

Source value: 86.3
Score: 86.3
Metric: percent
Unit: %
Category: knowledge
Direction: Higher is better
Source type: vendor-reported
Evaluator: official model card/release page
Status: active
Retrieved at: 2026-09-08
Methodology: Vendor-reported result. official page; asterisk row in-house

Source official model card/release page

Page 1 of 2

Predecessor and successor

Sources and data date

Type	Evidence and data date
Speed	Upstage. Output speed after the first streaming chunk, not total response time. Retrieved September 8, 2026. Source: Artificial Analysis · Artificial Analysis (retrieved September 8, 2026)Speed measurement methodology
Price	Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs. Source: Pricing · API pricing
Input	Documented modalities. Retrieved September 8, 2026. Source: Model documentation · Model metadata (retrieved September 8, 2026) · Upstage Solar Pro 4 announcement · Editorial description reviewed October 2, 2026
Output	Documented modalities. Retrieved September 8, 2026. Source: Model documentation · Model metadata (retrieved September 8, 2026) · Upstage Solar Pro 4 announcement · Editorial description reviewed October 2, 2026
Research status	Research date September 8, 2026. 2 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Not found in the inspected sources	parameters-billions Original wording No exact value located in the checked primary model card/API documentation.
Not found in the inspected sources	active-parameters-billions Original wording No exact value located in the checked primary model card/API documentation.
Not found in the inspected sources	knowledge-cutoff Original wording No exact value located in the checked primary model card/API documentation.
Additional source	OpenRouter: Popularity index (derived), CC BY 4.0
Additional source	Vercel AI Gateway: Popularity index (derived), CC BY 4.0
Additional source	Creative Commons Attribution 4.0

Solar Pro 4

Specifications and access

Measurements without matching peer values

Our tests

Pelican on a bicycle

Popularity

Published prices

Predecessor and successor

Sources and data date