Model comparison

Gemma 2 27B vs Solar Pro4

Solar Pro4 is the stronger model overall, scoring 42.1 to 29.4 on the Noometry Index.

Last verified . 17 shared benchmarks.

Gemma 2 27B Google

29.4

Rank #312 Confirmed

Solar Pro4 Upstage

42.1

Rank #121 Confirmed

Summary

  • They share 17 benchmarks with published results for both. Gemma 2 27B scores higher in 0 categories and Solar Pro4 in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where Solar Pro4 leads 38.8 to 10.7.
  • Solar Pro4 is cheaper at $0.30 / $1.20 per million input/output tokens, against $0.65 / $0.65 for Gemma 2 27B.
  • Solar Pro4 accepts more context: 524K tokens versus 8K.
  • Gemma 2 27B has downloadable open weights; the other is API-only.

Side by side

Gemma 2 27B and Solar Pro4 specifications
Gemma 2 27BSolar Pro4
ProviderGoogleUpstage
Noometry Index29.442.1
Released2024-06-242026-08-06
WeightsOpenProprietary
Context window8K524K
Max output2K131K
Input $ / M tokens$0.65$0.30
Output $ / M tokens$0.65$1.20
Results tracked3418

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Solar Pro4 leads

Gemma 2 27B: 34.1 (#246), Solar Pro4: 40.1 (#149)

Coding benchmarks
BenchmarkGemma 2 27BSolar Pro4
LMArena Coding12111437
LMArena WebDev—1371
BigCodeBench Instruct42.8%—
LiveBench Coding36%—
BigCodeBench Complete52.5%—

Reasoning Solar Pro4 leads

Gemma 2 27B: 15.3 (#315), Solar Pro4: 28.5 (#104)

Reasoning benchmarks
BenchmarkGemma 2 27BSolar Pro4
LMArena Hard Prompts11981399
LiveBench Reasoning28.1%—
DTBench48%—
LiveBench Data Analysis47.9%—
LMCA7.1%—
Epoch Capabilities Index122.08—
LiveBench38.2%—

Math Solar Pro4 leads

Gemma 2 27B: 10.7 (#311), Solar Pro4: 38.8 (#128)

Math benchmarks
BenchmarkGemma 2 27BSolar Pro4
LMArena Math12121416
OTIS Mock AIME 2024-20251.4%—
LiveBench Math26.5%—
MATH Level 527.9%—

Knowledge Solar Pro4 leads

Gemma 2 27B: 19.0 (#280), Solar Pro4: 39.8 (#129)

Knowledge benchmarks
BenchmarkGemma 2 27BSolar Pro4
LMArena Expert11721427
GPQA Diamond36.5%—
Confabulations27.1%—
MMLU75.7%—

Multilingual Solar Pro4 leads

Gemma 2 27B: 38.6 (#226), Solar Pro4: 48.7 (#139)

Multilingual benchmarks
BenchmarkGemma 2 27BSolar Pro4
LMArena Non-English12171361
LMArena Chinese12211415
LMArena French12471397
LMArena German12091364
LMArena Japanese11751309
LMArena Korean11741382
LMArena Russian12341360
LMArena Spanish12281401

Instruction Following Solar Pro4 leads

Gemma 2 27B: 60.5 (#249), Solar Pro4: 72.7 (#132)

Instruction Following benchmarks
BenchmarkGemma 2 27BSolar Pro4
LMArena Instruction Following12061377
LiveBench Instruction Following58.1%—

Long Context Solar Pro4 leads

Gemma 2 27B: 37.3 (#218), Solar Pro4: 42.1 (#130)

Long Context benchmarks
BenchmarkGemma 2 27BSolar Pro4
LMArena Longer Query12311381

Writing & Preference Solar Pro4 leads

Gemma 2 27B: 44.2 (#225), Solar Pro4: 56.5 (#138)

Writing & Preference benchmarks
BenchmarkGemma 2 27BSolar Pro4
LMArena Text12311386
LMArena Creative Writing12411316
LMArena Multi-Turn12241385
LiveBench Language32.6%—

Frequently asked questions

Is Gemma 2 27B better than Solar Pro4?

Solar Pro4 is the stronger model overall, scoring 42.1 to 29.4 on the Noometry Index.

Which is cheaper, Gemma 2 27B or Solar Pro4?

Solar Pro4 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; Gemma 2 27B lists at $0.65 and $0.65.

Is Gemma 2 27B or Solar Pro4 better for coding?

Solar Pro4 scores higher on coding benchmarks: 40.1 versus 34.1 in the Noometry coding category.

Which has the bigger context window?

Solar Pro4 does, with 524K tokens against 8K.

How many benchmarks do Gemma 2 27B and Solar Pro4 share?

17 benchmarks have published results for both models. Gemma 2 27B has 34 scored results on Noometry and Solar Pro4 has 18.

Related comparisons

Go deeper