Upstage model product
Solar Pro 4
A reasoning-focused agentic text-to-text LLM from Upstage for multi-step tasks.
Updated Oct 23, 2026. Default version: Solar Pro 4
LLMBoard Score55.1Solar Pro 4
Coverage60%9 benchmark families
Context window524.3KTokens
Official input price$0.30Upstage API
Solar Pro 4 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Solar Pro 4 Benchmark Results
Benchmark scores for Solar Pro 4.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkLiveCodeBench | Score87.80% | Rank05 | Participants76 | Percentile94.67% | EvidenceC | Evaluated |
| BenchmarkAA-LCR | Score71.00% | Rank07 | Participants27 | Percentile76.92% | EvidenceC | Evaluated |
| BenchmarkAIME 2026 | Score95.30% | Rank09 | Participants27 | Percentile69.23% | EvidenceC | Evaluated |
| BenchmarkTau3 Banking | Score23.00% | Rank09 | Participants16 | Percentile46.67% | EvidenceC | Evaluated |
| BenchmarkMMLU-Pro | Score86.30% | Rank13 | Participants142 | Percentile91.49% | EvidenceC | Evaluated |
| BenchmarkGPQA | Score89.00% | Rank37 | Participants256 | Percentile85.88% | EvidenceC | Evaluated |
| BenchmarkTerminal-Bench 2.1 | Score57.00% | Rank39 | Participants48 | Percentile19.15% | EvidenceC | Evaluated |
| BenchmarkLM Arena Agent Tool Hallucination | Score-2.44% | Rank40 | Participants44 | Percentile9.30% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Steerability | Score-14.53% | Rank41 | Participants44 | Percentile6.98% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Praise Complaint | Score-21.26% | Rank42 | Participants44 | Percentile4.65% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Leaderboard | Score-19.61% | Rank44 | Participants44 | Percentile0.00% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Bash Recovery Steps | Score-25.56% | Rank44 | Participants44 | Percentile0.00% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Task Outcome Explicit | Score-27.77% | Rank44 | Participants44 | Percentile0.00% | EvidenceA | Evaluated |
| BenchmarkAA GPQA Diamond | Score89.09% | Rank44 | Participants199 | Percentile78.28% | EvidenceB | EvaluatedN/A |
| BenchmarkLM Arena Agent Bash Reliability | Score-44.51% | Rank44 | Participants44 | Percentile0.00% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Bash Reliability No Prereq | Score-51.51% | Rank45 | Participants45 | Percentile0.00% | EvidenceA | Evaluated |
| BenchmarkAA CritPt | Score5.43% | Rank48 | Participants200 | Percentile76.38% | EvidenceB | EvaluatedN/A |
| BenchmarkAA SciCode Subtasks | Score44.56% | Rank49 | Participants89 | Percentile45.45% | EvidenceB | EvaluatedN/A |
| BenchmarkAA HLE Text No Tools | Score29.24% | Rank53 | Participants200 | Percentile73.87% | EvidenceB | EvaluatedN/A |
| BenchmarkBrowseComp | Score49.20% | Rank56 | Participants70 | Percentile20.29% | EvidenceC | Evaluated |
| BenchmarkAA LCR v1.1 | Score74.00% | Rank67 | Participants193 | Percentile65.63% | EvidenceB | EvaluatedN/A |
| BenchmarkSWE-Bench Verified | Score70.60% | Rank68 | Participants120 | Percentile43.70% | EvidenceC | Evaluated |
| BenchmarkLM Arena Webdev | Score1,370.68 rating | Rank78 | Participants107 | Percentile27.36% | EvidenceA | Evaluated |
| BenchmarkLM Arena Text Factuality | Score1,413.15 rating | Rank109 | Participants131 | Percentile16.92% | EvidenceA | Evaluated |
| BenchmarkLM Arena Text | Score1,387.91 rating | Rank121 | Participants218 | Percentile44.70% | EvidenceA | Evaluated |
| BenchmarkLM Arena Text Style Control | Score1,387.60 rating | Rank132 | Participants218 | Percentile39.63% | EvidenceA | Evaluated |
| BenchmarkAA Omniscience Accuracy | Score18.93% | Rank133 | Participants201 | Percentile34.00% | EvidenceB | EvaluatedN/A |
Solar Pro 4 Arena Results
Preference and agent-evaluation results for the default version.
Sort by
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenaagent tool hallucination | Categoryoverall | Rank40 | Rating / score-0.02 | VotesN/A | Observations442.4K | Result date |
| Arenaagent steerability | Categoryoverall | Rank41 | Rating / score-0.15 | VotesN/A | Observations23.8K | Result date |
| Arenaagent praise complaint | Categoryoverall | Rank42 | Rating / score-0.21 | VotesN/A | Observations2.1K | Result date |
| Arenaagent | Categoryoverall | Rank44 | Rating / score-0.20 | VotesN/A | Observations485.5K | Result date |
| Arenaagent bash recovery steps | Categoryoverall | Rank44 | Rating / score-0.26 | VotesN/A | Observations23.4K | Result date |
| Arenaagent task outcome explicit | Categoryoverall | Rank44 | Rating / score-0.28 | VotesN/A | Observations4.6K | Result date |
| Arenaagent bash reliability | Categoryoverall | Rank44 | Rating / score-0.45 | VotesN/A | Observations12.6K | Result date |
| Arenaagent bash reliability no prereq | Categoryoverall | Rank45 | Rating / score-0.52 | VotesN/A | Observations6.5K | Result date |
| Arenatext factuality | Categoryindustry mathematical | Rank62 | Rating / score1,435.35 | Votes1,068 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryexpert | Rank71 | Rating / score1,460.68 | Votes2,317 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev-react | Rank74 | Rating / score1,363.23 | Votes2,953 | ObservationsN/A | Result date |
| Arenatext | Categorykorean | Rank75 | Rating / score1,383.15 | Votes367 | ObservationsN/A | Result date |
| Arenatext factuality | Categorymath | Rank75 | Rating / score1,421.32 | Votes924 | ObservationsN/A | Result date |
| Arenatext factuality | Categorychinese | Rank78 | Rating / score1,457.81 | Votes1,481 | ObservationsN/A | Result date |
| Arenawebdev | Categoryoverall | Rank78 | Rating / score1,370.68 | Votes4,655 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev | Rank78 | Rating / score1,370.68 | Votes4,655 | ObservationsN/A | Result date |
| Arenawebdev | Categorywebdev-html | Rank78 | Rating / score1,370.63 | Votes813 | ObservationsN/A | Result date |
| Arenatext style control | Categorykorean | Rank83 | Rating / score1,374.94 | Votes367 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry legal and government | Rank85 | Rating / score1,430.08 | Votes1,626 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry medicine and healthcare | Rank88 | Rating / score1,422.10 | Votes1,260 | ObservationsN/A | Result date |
| Arenatext factuality | Categorycoding | Rank90 | Rating / score1,474.15 | Votes5,836 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryhard prompts english | Rank90 | Rating / score1,458.84 | Votes5,860 | ObservationsN/A | Result date |
| Arenatext | Categorymath | Rank92 | Rating / score1,420.86 | Votes1,181 | ObservationsN/A | Result date |
| Arenatext | Categoryexpert | Rank93 | Rating / score1,427.99 | Votes2,674 | ObservationsN/A | Result date |
| Arenatext | Categorycoding | Rank95 | Rating / score1,437.51 | Votes5,940 | ObservationsN/A | Result date |
| Arenatext | Categoryindustry mathematical | Rank95 | Rating / score1,427.01 | Votes1,360 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry software and it services | Rank95 | Rating / score1,460.31 | Votes8,785 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry life and physical and social science | Rank98 | Rating / score1,437.53 | Votes3,521 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryenglish | Rank99 | Rating / score1,433.28 | Votes9,217 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryrussian | Rank100 | Rating / score1,398.06 | Votes2,566 | ObservationsN/A | Result date |
Solar Pro 4 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $0.30 input, $1.2 output per 1M
- Official provider
- Upstage
- Lowest third-party
- From $0.03 input, $0.12 output per 1M via NanoGPT
- Tracked offerings
- 6
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderNanoGPT | Provider model IDupstage/solar-pro4 | Regionglobal | Input / 1M$0.03 | Output / 1M$0.12 | Context524.3K | Updated |
| ProviderAI/ML API | Provider model IDupstage/solar-pro4 | Regionglobal | Input / 1M$0.0317 | Output / 1M$0.127 | Context524.3K | Updated |
| ProviderOpenRouter | Provider model IDupstage/solar-pro4 | Regionglobal | Input / 1M$0.15 | Output / 1M$0.60 | Context524.3K | Updated |
| ProviderLLMTR | Provider model IDupstage/solar-pro4 | Regionglobal | Input / 1M$0.15 | Output / 1M$0.60 | Context524.3K | Updated |
| ProviderUpstage | Provider model IDsolar-pro4 | Regionglobal | Input / 1M$0.30 | Output / 1M$1.2 | Context524.3K | Updated |
| ProviderKilo Gateway | Provider model IDupstage/solar-pro4 | Regionglobal | Input / 1M$0.30 | Output / 1M$1.2 | Context524.3K | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Solar Pro 4 Runtime Performance
Provider-specific output speed and catalog latency for Solar Pro 4. Runtime does not affect the capability score.
No runtime data
No provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Solar Pro 4 Specifications
Technical details for the model's default version.
- Version
- Solar Pro 4
- Released
- Aug 6, 2026
- Knowledge cutoff
- Unknown
- Parameters
- N/A
- Context window
- 524.3K
- Max output
- 131.1K
- Inputs
- text
- Outputs
- text
- Open weights
- No
- License
- Proprietary
Solar Pro 4 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionSolar Pro 4 | Released | LLMBoard55.14 | ParametersN/A | Context524.3K | Max output131.1K | Open weightsNo | LicenseProprietary |
Models similar to Solar Pro 4
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
UP
Solar Mini 4
Upstage
AN
Claude Sonnet 4.5
Anthropic
AC
Qwen3.5 27B
Alibaba Cloud / Qwen Team
BY
Seed 1.8
ByteDance
XA
Grok 4
xAI
DE
DeepSeek-V3.2
DeepSeek
What is Solar Pro 4?
Key information about Solar Pro 4 and its available data.
Solar Pro 4 is Upstage's proprietary reasoning-focused agentic LLM for text-to-text tasks, supporting English, Korean, and Japanese. It has a ~512K context window, up to 128K output tokens, and adjustable reasoning effort.
Data as of 2026-10-22.
FAQ
Common questions about Solar Pro 4.
When was Solar Pro 4 released?
Solar Pro 4's default version was released on Aug 6, 2026.
How much does Solar Pro 4 cost?
Solar Pro 4's official API price is $0.30 per million input tokens and $1.2 per million output tokens via Upstage. The lowest tracked third-party offer starts at $0.03 input and $0.12 output via NanoGPT.
Who created Solar Pro 4?
Solar Pro 4 was created by Upstage.
What is the context window for Solar Pro 4?
The default version has a 524.3K token context window.
Is Solar Pro 4 open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer Solar Pro 4?
6 provider offerings are linked to the default version.