Noometry

Model comparison

GPT-6 Luna vs phi-3-medium 14B

GPT-6 Luna is the stronger model overall, scoring 53.3 to 29.7 on the Noometry Index.

Last verified . 2 shared benchmarks.

GPT-6 Luna OpenAI

53.3

Rank #36 Confirmed

phi-3-medium 14B Microsoft

29.7

Rank #306 Reported

Summary

  • They share 2 benchmarks with published results for both. GPT-6 Luna scores higher in 3 categories and phi-3-medium 14B in 0 categories; 3 gaps are clear of the uncertainty.
  • The widest gap is in math, where GPT-6 Luna leads 76.1 to 27.3.
  • The biggest single-benchmark swing is GPQA Diamond: 90.5% for GPT-6 Luna and 27.6% for phi-3-medium 14B.
  • phi-3-medium 14B has downloadable open weights; the other is API-only.

Side by side

GPT-6 Lunaphi-3-medium 14B
ProviderOpenAIMicrosoft
Noometry Index53.329.7
Released2026-09-222024-04-23
WeightsProprietaryOpen
Context window1.05M—
Max output128K—
Input $ / M tokens$0.10—
Output $ / M tokens$0.50—
Results tracked4213

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

Category by category

Coding GPT-6 Luna leads

GPT-6 Luna: 55.5 (#25), phi-3-medium 14B: 36.8 (#201)

BenchmarkGPT-6 Lunaphi-3-medium 14B
DeepSWE66.6%—
FrontierCode42.4%—
LMArena WebDev1581—
SciCode54.6%—
BigCodeBench Instruct—37.6%
LMArena Coding1439—
BigCodeBench Complete—48.7%
ALE-Bench1,577—

Agentic & Tool Use Not comparable

GPT-6 Luna: 33.3 (#54), phi-3-medium 14B: —

BenchmarkGPT-6 Lunaphi-3-medium 14B
APEX-Agents44.3%—
GDP.pdf23%—

Reasoning Not comparable

GPT-6 Luna: 48.2 (#41), phi-3-medium 14B: —

BenchmarkGPT-6 Lunaphi-3-medium 14B
Epoch Capabilities Index156.28121.23
ARC-AGI-259.3%—
NYT Connections (extended)68.7%—
ARC-AGI-186.7%—
CritPt19.4%—
Chess Puzzles31%—
LMArena Hard Prompts1411—
Mystery Game Puzzles7%—
DTBench90.1%—
LMCA44.5%—
Adversarial NLI—55.8%
BIG-Bench Hard—81.4%
HellaSwag—82.4%
WinoGrande—81.5%

Math GPT-6 Luna leads

GPT-6 Luna: 76.1 (#15), phi-3-medium 14B: 27.3 (#250)

BenchmarkGPT-6 Lunaphi-3-medium 14B
FrontierMath (Tiers 1-3)78.9%—
FrontierMath Tier 456.1%—
OTIS Mock AIME 2024-202598.9%—
ProofBench64%—
LMArena Math1416—
MATH Level 5—17.6%

Knowledge GPT-6 Luna leads

GPT-6 Luna: 57.0 (#41), phi-3-medium 14B: 9.1 (#306)

BenchmarkGPT-6 Lunaphi-3-medium 14B
GPQA Diamond90.5%27.6%
SimpleQA Verified41.4%—
LMArena Expert1444—
ARC (AI2) Challenge—91.6%
MMLU—78%
OpenBookQA—87.4%
TriviaQA—73.9%

Multimodal Not comparable

GPT-6 Luna: 42.4 (#30), phi-3-medium 14B: —

BenchmarkGPT-6 Lunaphi-3-medium 14B
LMArena Vision1217—
Blueprint-Bench 231.2%—
Furniture Assembly44.2%—

Multilingual Not comparable

GPT-6 Luna: 50.5 (#117), phi-3-medium 14B: —

BenchmarkGPT-6 Lunaphi-3-medium 14B
LMArena Non-English1386—
LMArena Chinese1433—
LMArena French1420—
LMArena German1369—
LMArena Japanese1369—
LMArena Korean1360—
LMArena Russian1394—
LMArena Spanish1393—

Instruction Following Not comparable

GPT-6 Luna: 74.3 (#99), phi-3-medium 14B: —

BenchmarkGPT-6 Lunaphi-3-medium 14B
LMArena Instruction Following1409—

Long Context Not comparable

GPT-6 Luna: 43.0 (#111), phi-3-medium 14B: —

BenchmarkGPT-6 Lunaphi-3-medium 14B
LMArena Longer Query1409—

Writing & Preference Not comparable

GPT-6 Luna: 58.3 (#119), phi-3-medium 14B: —

BenchmarkGPT-6 Lunaphi-3-medium 14B
LMArena Text1391—
LMArena Creative Writing1363—
LMArena Multi-Turn1396—

Frequently asked questions

Is GPT-6 Luna better than phi-3-medium 14B?

GPT-6 Luna is the stronger model overall, scoring 53.3 to 29.7 on the Noometry Index.

Is GPT-6 Luna or phi-3-medium 14B better for coding?

GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 36.8 in the Noometry coding category.

How many benchmarks do GPT-6 Luna and phi-3-medium 14B share?

2 benchmarks have published results for both models. GPT-6 Luna has 42 scored results on Noometry and phi-3-medium 14B has 13.

Related comparisons

Go deeper