Noometry

Model comparison

GPT-5.6 Luna vs Qwen3.8 Max

Qwen3.8 Max is the stronger model overall, scoring 56.8 to 54.6 on the Noometry Index. GPT-5.6 Luna costs 6.7× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.

Last verified . 37 shared benchmarks.

GPT-5.6 Luna OpenAI

54.6

Rank #30 Confirmed

Qwen3.8 Max Alibaba (Qwen)

56.8

Rank #22 Confirmed

Summary

  • They share 37 benchmarks with published results for both. GPT-5.6 Luna scores higher in 4 categories and Qwen3.8 Max in 6 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in agentic & tool use, where Qwen3.8 Max leads 45.4 to 34.4.
  • The biggest single-benchmark swing is Furniture Assembly: 42.5% for GPT-5.6 Luna and 20% for Qwen3.8 Max.
  • GPT-5.6 Luna is cheaper at $0.20 / $1.20 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
  • GPT-5.6 Luna accepts more context: 1.05M tokens versus 1M.

Side by side

GPT-5.6 LunaQwen3.8 Max
ProviderOpenAIAlibaba (Qwen)
Noometry Index54.656.8
Released2026-07-092026-08-02
WeightsProprietaryProprietary
Context window1.05M1M
Max output128K131K
Input $ / M tokens$0.20$2
Output $ / M tokens$1.20$6
Results tracked5239

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

Category by category

Coding Too close to call

GPT-5.6 Luna: 54.5 (#28), Qwen3.8 Max: 53.5 (#29)

BenchmarkGPT-5.6 LunaQwen3.8 Max
DeepSWE67.2%57.5%
LMArena WebDev15191674
SciCode53.6%53.2%
LMArena Coding14661502
FrontierCode39.8%—
CursorBench35.9%—
FrontierSWE—17.8%
WeirdML60.9%—
ALE-Bench1,667—

Agentic & Tool Use Qwen3.8 Max leads

GPT-5.6 Luna: 34.4 (#45), Qwen3.8 Max: 45.4 (#14)

BenchmarkGPT-5.6 LunaQwen3.8 Max
APEX-Agents43%63.3%
GDP.pdf22.7%23.2%
τ²-bench Banking—55.1%
BALROG45.6%—
Vending-Bench 24,095—

Reasoning Qwen3.8 Max leads

GPT-5.6 Luna: 47.6 (#43), Qwen3.8 Max: 54.4 (#26)

BenchmarkGPT-5.6 LunaQwen3.8 Max
NYT Connections (extended)69.4%88.3%
CritPt20.6%20%
Chess Puzzles40%40%
LMArena Hard Prompts14511496
Mystery Game Puzzles21%38%
DTBench89.1%92%
LMCA48.5%46.2%
Epoch Capabilities Index156.39156.41
ARC-AGI-259.5%—
SimpleBench46.8%—
Kagi LLM Benchmark49.1%—
ARC-AGI-188%—
Surface Evolver Bench61.9%—

Math GPT-5.6 Luna leads

GPT-5.6 Luna: 77.7 (#14), Qwen3.8 Max: 73.2 (#20)

BenchmarkGPT-5.6 LunaQwen3.8 Max
FrontierMath (Tiers 1-3)82.1%74.7%
FrontierMath Tier 461%46.3%
OTIS Mock AIME 2024-202598.3%100%
ProofBench60%58%
LMArena Math14581499

Knowledge Qwen3.8 Max leads

GPT-5.6 Luna: 58.5 (#34), Qwen3.8 Max: 61.7 (#27)

BenchmarkGPT-5.6 LunaQwen3.8 Max
GPQA Diamond91.6%92.7%
SimpleQA Verified41%47.3%
LMArena Expert14781507

Multimodal GPT-5.6 Luna leads

GPT-5.6 Luna: 42.7 (#28), Qwen3.8 Max: 37.2 (#75)

BenchmarkGPT-5.6 LunaQwen3.8 Max
LMArena Vision12581314
Furniture Assembly42.5%20%
Blueprint-Bench 222.6%—
LMArena Document1457—

Multilingual Qwen3.8 Max leads

GPT-5.6 Luna: 52.8 (#78), Qwen3.8 Max: 56.7 (#18)

BenchmarkGPT-5.6 LunaQwen3.8 Max
LMArena Non-English14171472
LMArena Chinese14701538
LMArena French14561503
LMArena German14541483
LMArena Japanese14111467
LMArena Korean14151461
LMArena Russian14281481
LMArena Spanish14481492

Instruction Following Qwen3.8 Max leads

GPT-5.6 Luna: 75.6 (#57), Qwen3.8 Max: 77.6 (#17)

BenchmarkGPT-5.6 LunaQwen3.8 Max
LMArena Instruction Following14371479

Long Context Qwen3.8 Max leads

GPT-5.6 Luna: 43.9 (#82), Qwen3.8 Max: 45.6 (#31)

BenchmarkGPT-5.6 LunaQwen3.8 Max
LMArena Longer Query14361489

Writing & Preference Too close to call

GPT-5.6 Luna: 68.0 (#29), Qwen3.8 Max: 67.1 (#30)

BenchmarkGPT-5.6 LunaQwen3.8 Max
LMArena Text14311483
LMArena Creative Writing13961479
LMArena Multi-Turn14341489
EQ-Bench Creative Writing1829—
EQ-Bench 41156—

Frequently asked questions

Is GPT-5.6 Luna better than Qwen3.8 Max?

Qwen3.8 Max is the stronger model overall, scoring 56.8 to 54.6 on the Noometry Index. GPT-5.6 Luna costs 6.7× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.

Which is cheaper, GPT-5.6 Luna or Qwen3.8 Max?

GPT-5.6 Luna is cheaper. It lists at $0.20 per million input tokens and $1.20 per million output tokens; Qwen3.8 Max lists at $2 and $6.

Is GPT-5.6 Luna or Qwen3.8 Max better for coding?

They score almost the same on coding (54.5 vs 53.5); test both on your own repository before choosing.

Which has the bigger context window?

GPT-5.6 Luna does, with 1.05M tokens against 1M.

How many benchmarks do GPT-5.6 Luna and Qwen3.8 Max share?

37 benchmarks have published results for both models. GPT-5.6 Luna has 52 scored results on Noometry and Qwen3.8 Max has 39.

Related comparisons

Go deeper