SShogo Bench

GPT-5.6 Luna vs Muse Spark 1.3

Muse Spark 1.3 leads GPT-5.6 Luna by 3.3 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. GPT-5.6 Luna wins 4 and Muse Spark 1.3 wins 11 of 15 shared benchmarks. Muse Spark 1.3 is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarGPT-5.6 LunaMuse Spark 1.3
Agentic coding5960
Long-horizon autonomy6164
Knowledge-work agents3945
Computer use4043
Security & context learning5153

Shared benchmarks

BenchmarkGPT-5.6 LunaMuse Spark 1.3
APEX-Agentsindex43.0%58.6%
CursorBenchindex35.9%41.6%
Chess Puzzles43.0%37.0%
CritPt20.6%26.0%
DTBench89.1%96.5%
FrontierMath Tier 461.0%46.3%
FrontierMath Tiers 1-382.1%74.4%
LMCA48.5%53.9%
MirrorCode21.1%54.4%
OTIS Mock AIME98.3%99.2%
ProofBench60.0%58.0%
SciCode53.6%59.7%
SimpleBench46.8%81.8%
WebDev Arena15191652
WeirdML v35.2%6.7%

Specs and price

GPT-5.6 LunaMuse Spark 1.3
Agent Index51.4 (#34)54.7 (#28)
Agent $/M$0.10$0.030
Input / output $/M$0.20 / $1.20$0.10 / $0.20
Context1.1M1M
ReleasedJul 9, 2026Sep 2, 2026
WeightsClosedClosed
API hosts52

Don’t pick one. Route.

Shogo can run GPT-5.6 Luna and Muse Spark 1.3 side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo