Noometry

Model comparison

Phi 3 Small 8k Instruct vs Qwen3.8 Max

Qwen3.8 Max is the stronger model overall, scoring 56.8 to 29.3 on the Noometry Index.

Last verified . 17 shared benchmarks.

Phi 3 Small 8k Instruct Microsoft

29.3

Rank #314 Confirmed

Qwen3.8 Max Alibaba (Qwen)

56.8

Rank #22 Confirmed

Summary

  • They share 17 benchmarks with published results for both. Phi 3 Small 8k Instruct scores higher in 0 categories and Qwen3.8 Max in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where Qwen3.8 Max leads 73.2 to 27.6.
  • Phi 3 Small 8k Instruct has downloadable open weights; the other is API-only.

Side by side

Phi 3 Small 8k InstructQwen3.8 Max
ProviderMicrosoftAlibaba (Qwen)
Noometry Index29.356.8
Released2024-04-232026-08-02
WeightsOpenProprietary
Context window—1M
Max output—131K
Input $ / M tokens—$2
Output $ / M tokens—$6
Results tracked3239

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

Category by category

Coding Qwen3.8 Max leads

Phi 3 Small 8k Instruct: 27.9 (#318), Qwen3.8 Max: 53.5 (#29)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Coding11011502
DeepSWE—57.5%
LMArena WebDev—1674
FrontierSWE—17.8%
SciCode—53.2%
LiveBench Coding20.3%—

Agentic & Tool Use Not comparable

Phi 3 Small 8k Instruct: —, Qwen3.8 Max: 45.4 (#14)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
APEX-Agents—63.3%
τ²-bench Banking—55.1%
GDP.pdf—23.2%

Reasoning Qwen3.8 Max leads

Phi 3 Small 8k Instruct: 14.8 (#323), Qwen3.8 Max: 54.4 (#26)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Hard Prompts11001496
NYT Connections (extended)—88.3%
CritPt—20%
Chess Puzzles—40%
LiveBench Reasoning15.9%—
Mystery Game Puzzles—38%
DTBench—92%
LiveBench Data Analysis30.3%—
LMCA—46.2%
Adversarial NLI58.1%—
BIG-Bench Hard79.1%—
Epoch Capabilities Index—156.41
HellaSwag77%—
LiveBench24%—
WinoGrande81.5%—

Math Qwen3.8 Max leads

Phi 3 Small 8k Instruct: 27.6 (#248), Qwen3.8 Max: 73.2 (#20)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Math11511499
FrontierMath (Tiers 1-3)—74.7%
FrontierMath Tier 4—46.3%
OTIS Mock AIME 2024-2025—100%
ProofBench—58%
LiveBench Math17.6%—

Knowledge Qwen3.8 Max leads

Phi 3 Small 8k Instruct: 29.1 (#240), Qwen3.8 Max: 61.7 (#27)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Expert10671507
GPQA Diamond—92.7%
SimpleQA Verified—47.3%
ARC (AI2) Challenge90.7%—
MMLU75.7%—
OpenBookQA88%—
TriviaQA58.1%—

Multimodal Not comparable

Phi 3 Small 8k Instruct: —, Qwen3.8 Max: 37.2 (#75)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Vision—1314
Furniture Assembly—20%

Multilingual Qwen3.8 Max leads

Phi 3 Small 8k Instruct: 28.5 (#272), Qwen3.8 Max: 56.7 (#18)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Non-English10581472
LMArena Chinese10611538
LMArena French11351503
LMArena German10801483
LMArena Japanese9661467
LMArena Korean8941461
LMArena Russian11111481
LMArena Spanish11111492

Instruction Following Qwen3.8 Max leads

Phi 3 Small 8k Instruct: 51.9 (#292), Qwen3.8 Max: 77.6 (#17)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Instruction Following10871479
LiveBench Instruction Following47.2%—

Long Context Qwen3.8 Max leads

Phi 3 Small 8k Instruct: 33.0 (#267), Qwen3.8 Max: 45.6 (#31)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Longer Query10881489

Writing & Preference Qwen3.8 Max leads

Phi 3 Small 8k Instruct: 31.1 (#284), Qwen3.8 Max: 67.1 (#30)

BenchmarkPhi 3 Small 8k InstructQwen3.8 Max
LMArena Text11101483
LMArena Creative Writing10831479
LMArena Multi-Turn10681489
LiveBench Language12.9%—

Frequently asked questions

Is Phi 3 Small 8k Instruct better than Qwen3.8 Max?

Qwen3.8 Max is the stronger model overall, scoring 56.8 to 29.3 on the Noometry Index.

Is Phi 3 Small 8k Instruct or Qwen3.8 Max better for coding?

Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 27.9 in the Noometry coding category.

How many benchmarks do Phi 3 Small 8k Instruct and Qwen3.8 Max share?

17 benchmarks have published results for both models. Phi 3 Small 8k Instruct has 32 scored results on Noometry and Qwen3.8 Max has 39.

Related comparisons

Go deeper