Solar Pro 4

Solar Pro 4 is a reasoning model from Upstage in the Solar family. 17 benchmarks count toward its score, in 5 categories.

availableShows if the model has enough results for an index.
IndexOverall score out of 100.53.7 ±5.9
CoverageShare of the index weight with results.80%
SpeedOutput tokens per second.38/s
Input / 1MUS dollars per 1M input tokens.$0.09
Output / 1MUS dollars per 1M output tokens.$0.36
ContextMaximum tokens in one request.524K
EloLMArena rating and rank.1377 (#161)

The index is a score out of 100. The ± range shows how much it can change.

2,420 votes. Elo shows what people prefer. It does not change the score.

CapabilitiesScore per category, out of 100.

Out of 100
AgenticMulti-step tasks with tools.
50.7
CodingCode writing and repair.
57.7
ReasoningLogic problems and puzzles.
53.5
MultimodalTasks with images and text.
N/A
KnowledgeFacts and expert knowledge.
54.6
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
N/A
MathMath problems.
54.4

Results

17 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result as a score out of 100.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
AIME 2026Math95.3%54.4Qwen
Artificial Analysis GPQA DiamondKnowledge89.1%59.9Artificial Analysis
GPQA DiamondKnowledge89.0%60.4David Rein et al.
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for CodeCoding87.8%64.1Naman Jain et al.
Massive Multitask Language Understanding ProfessionalKnowledge86.3%56.5Yubo Wang et al.
Artificial Analysis Long Context ReasoningReasoning71.0%57.4Artificial Analysis
Software Engineering Benchmark VerifiedCoding70.6%54.2Carlos E. Jimenez et al.
MCP AtlasAgentic61.4%54.4OpenAI
Terminal-Bench 2.1 (provider run)Agentic57.0%57.7DeepSeek-AI
Terminal-Bench 2.1 (provider run)Agentic57.0%57.7DeepSeek-AI
BrowseCompAgentic49.2%41.3OpenAI
Artificial Analysis SciCodeCoding44.6%54.8Artificial Analysis
Artificial Analysis Humanity's Last ExamKnowledge29.2%56.2Artificial Analysis
Artificial Analysis Intelligence IndexKnowledge28.1%57.6Artificial Analysis
Artificial Analysis Omniscience AccuracyKnowledge18.9%37.3Artificial Analysis
APEX-AgentsAgentic18.7%55.1Moonshot AI / APEX-Agents benchmark authors
Critical Physics TasksReasoning5.4%49.7Artificial Analysis
Agent Arena steerabilityAgentic-10.156.015 Sept 2026LMArena
Agent Arena task outcomeAgentic-20.144.715 Sept 2026LMArena
Agent Arena command recoveryAgentic-24.240.115 Sept 2026LMArena

17 benchmarks count, from 20 of 20 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsOpenRouter, collected directlyNo licence statedLMArena, collected directlyCC BY 4.0 (lmarena-ai/leaderboard-dataset on Hugging Face)

Same level, lower price

Qwen3.8-Flash-Next66.4 · Freedots3-note Preview65.5 · FreeApodex 1.1 Mini59.1 · Free

More from Upstage

Solar Open 255.0Solar Pro 339.1Solar Pro 227.9