Inference
EN
Upstage

Solar Open 2

Solar Open 2 is a reasoning model from Upstage in the Solar Open family. 12 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-07
RankedEnough results for an index.
IndexAll categories, out of 100.53.3±6.2Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

250.29B parameters
16-bit
526.6 GB
8-bit
280.2 GB
4-bit
160.3 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy needs more than 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.3 counted47.9149 / 218
CodingCode writing and repair.2 counted56.787 / 209
ReasoningLogic problems and puzzles.1 counted50.0—
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.3 counted54.2101 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted63.8—
MathMath problems.2 counted56.759 / 168

Results

12 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
LiveCodeBench v6Coding92.4%4 / 59
Massive Multitask Language Understanding ProfessionalKnowledge86.2%12 / 75
Harvard-MIT Mathematics Tournament February 2026Math93.9%5 / 27
Instruction Following BenchmarkInstruction80.0%12 / 57
AIME 2026Math95.7%11 / 34
Software Engineering Benchmark VerifiedCoding70.4%70 / 114
GPQA Diamond · BenchLMKnowledge86.3%53 / 84
Artificial Analysis Long Context ReasoningReasoning62.3%139 / 217
MCP AtlasAgentic58.2%38 / 52
Humanity's Last Exam without toolsKnowledge28.8%41 / 53
Terminal-Bench HardAgentic28.3%23 / 23
APEX-AgentsAgentic16.6%13 / 13

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index55.0 → 53.3
23 Oct 202612 Nov 2026

Questions

How good is Solar Open 2?
Its index is 53.3 out of 100. That is place 115 of 283 ranked models.
What is Solar Open 2 best at?
Math: 56.7 out of 100, place 59 of 168.
What is the context window of Solar Open 2?
1M tokens.
Is Solar Open 2 open source?
Yes. Its weights are published.
When was Solar Open 2 released?
July 2026.

Badge

Solar Open 2 on Inference360: place 115 of 283, index 53.3
HTML
<a href="https://inference360.ai/models/solar-open-2"><img src="https://inference360.ai/badges/solar-open-2.svg" alt="Solar Open 2 on Inference360: place 115 of 283, index 53.3" height="20"></a>
Markdown
[![Solar Open 2 on Inference360: place 115 of 283, index 53.3](https://inference360.ai/badges/solar-open-2.svg)](https://inference360.ai/models/solar-open-2)

Compare with

Muse Glimmer 30B53.0Solar Pro 452.1

More from Upstage

All models
Solar Pro 452.1Solar Mini 4UnrankedSolar Pro 339.2Solar Pro 2Unranked