Noometry

Model comparison

Gemini 3.6 Flash vs Qwen3.8 27B

Gemini 3.6 Flash is the stronger model overall, scoring 54.1 to 45.3 on the Noometry Index.

Last verified . 27 shared benchmarks.

Gemini 3.6 Flash Google

54.1

Rank #30 Confirmed

Qwen3.8 27B Alibaba (Qwen)

45.3

Rank #64 Confirmed

Summary

  • They share 27 benchmarks with published results for both. Gemini 3.6 Flash scores higher in 7 categories and Qwen3.8 27B in 2 categories; 7 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Gemini 3.6 Flash leads 58.9 to 31.5.
  • The biggest single-benchmark swing is NYT Connections (extended): 89% for Gemini 3.6 Flash and 54.5% for Qwen3.8 27B.
  • Qwen3.8 27B is cheaper at $0.99 / $1.49 per million input/output tokens, against $0.75 / $3.75 for Gemini 3.6 Flash.
  • Gemini 3.6 Flash accepts more context: 1.05M tokens versus 262K.
  • Qwen3.8 27B has downloadable open weights; the other is API-only.

Side by side

Gemini 3.6 FlashQwen3.8 27B
ProviderGoogleAlibaba (Qwen)
Noometry Index54.145.3
Released2026-07-212026-08-14
WeightsProprietaryOpen
Context window1.05M262K
Max output66K33K
Input $ / M tokens$0.75$0.99
Output $ / M tokens$3.75$1.49
Results tracked4628

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

Category by category

Coding Too close to call

Gemini 3.6 Flash: 49.6 (#45), Qwen3.8 27B: 50.3 (#42)

BenchmarkGemini 3.6 FlashQwen3.8 27B
LMArena WebDev15361592
SciCode52.7%46.6%
LMArena Coding14921487
DeepSWE46.7%—
FrontierCode34.4%—
WeirdML56.1%—
ALE-Bench715.52—

Agentic & Tool Use Not comparable

Gemini 3.6 Flash: 32.4 (#56), Qwen3.8 27B: —

BenchmarkGemini 3.6 FlashQwen3.8 27B
APEX-Agents46.9%—
GDP.pdf14%—

Reasoning Gemini 3.6 Flash leads

Gemini 3.6 Flash: 58.9 (#20), Qwen3.8 27B: 31.5 (#71)

BenchmarkGemini 3.6 FlashQwen3.8 27B
NYT Connections (extended)89%54.5%
CritPt10.6%5.4%
LMArena Hard Prompts14861459
DTBench95.5%88%
LMCA44.9%41.4%
Epoch Capabilities Index154.36149.38
ARC-AGI-260.4%—
ARC-AGI-191.2%—
Chess Puzzles43%—
Mystery Game Puzzles30%—
Surface Evolver Bench—45%

Math Gemini 3.6 Flash leads

Gemini 3.6 Flash: 57.0 (#33), Qwen3.8 27B: 35.7 (#139)

BenchmarkGemini 3.6 FlashQwen3.8 27B
ProofBench36%16%
LMArena Math15041453
FrontierMath (Tiers 1-3)58.9%—
FrontierMath Tier 422%—
MathArena Final-Answer Competitions70.8%—
OTIS Mock AIME 2024-202594.2%—

Knowledge Gemini 3.6 Flash leads

Gemini 3.6 Flash: 67.8 (#7), Qwen3.8 27B: 40.9 (#105)

BenchmarkGemini 3.6 FlashQwen3.8 27B
LMArena Expert14891482
GPQA Diamond94.1%—
SimpleQA Verified66.2%—

Multimodal Qwen3.8 27B leads

Gemini 3.6 Flash: 38.7 (#55), Qwen3.8 27B: 40.5 (#36)

BenchmarkGemini 3.6 FlashQwen3.8 27B
LMArena Vision12971270
Blueprint-Bench 231.2%—
Furniture Assembly23.3%—
LMArena Document1456—

Multilingual Gemini 3.6 Flash leads

Gemini 3.6 Flash: 56.8 (#17), Qwen3.8 27B: 54.0 (#54)

BenchmarkGemini 3.6 FlashQwen3.8 27B
LMArena Non-English14701431
LMArena Chinese15301504
LMArena French15011468
LMArena German14781435
LMArena Japanese14771392
LMArena Korean14351389
LMArena Russian14901421
LMArena Spanish14771444

Instruction Following Gemini 3.6 Flash leads

Gemini 3.6 Flash: 77.2 (#27), Qwen3.8 27B: 75.8 (#54)

BenchmarkGemini 3.6 FlashQwen3.8 27B
LMArena Instruction Following14671436

Long Context Too close to call

Gemini 3.6 Flash: 45.1 (#47), Qwen3.8 27B: 44.3 (#66)

BenchmarkGemini 3.6 FlashQwen3.8 27B
LMArena Longer Query14751448

Writing & Preference Gemini 3.6 Flash leads

Gemini 3.6 Flash: 68.3 (#26), Qwen3.8 27B: 65.9 (#41)

BenchmarkGemini 3.6 FlashQwen3.8 27B
LMArena Text14781441
LMArena Creative Writing14641382
EQ-Bench Creative Writing16041671
LMArena Multi-Turn14811443

Frequently asked questions

Is Gemini 3.6 Flash better than Qwen3.8 27B?

Gemini 3.6 Flash is the stronger model overall, scoring 54.1 to 45.3 on the Noometry Index.

Which is cheaper, Gemini 3.6 Flash or Qwen3.8 27B?

Qwen3.8 27B is cheaper. It lists at $0.99 per million input tokens and $1.49 per million output tokens; Gemini 3.6 Flash lists at $0.75 and $3.75.

Is Gemini 3.6 Flash or Qwen3.8 27B better for coding?

They score almost the same on coding (49.6 vs 50.3); test both on your own repository before choosing.

Which has the bigger context window?

Gemini 3.6 Flash does, with 1.05M tokens against 262K.

How many benchmarks do Gemini 3.6 Flash and Qwen3.8 27B share?

27 benchmarks have published results for both models. Gemini 3.6 Flash has 46 scored results on Noometry and Qwen3.8 27B has 28.

Related comparisons

Go deeper