SShogo Bench

GPT-5.6 Luna vs Kimi K3

Kimi K3 leads GPT-5.6 Luna by 3.5 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. GPT-5.6 Luna wins 4 and Kimi K3 wins 17 of 21 shared benchmarks. GPT-5.6 Luna is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarGPT-5.6 LunaKimi K3
Agentic coding5963
Long-horizon autonomy6165
Knowledge-work agents3942
Computer use4043
Security & context learning5153

Shared benchmarks

BenchmarkGPT-5.6 LunaKimi K3
APEX-Agentsindex43.0%47.8%
DeepSWEindex67.2%68.5%
Vending-Bench 2index$4,095$5,165
ARC-AGI88.0%94.5%
ARC-AGI-259.5%60.4%
Chess Puzzles43.0%37.0%
CritPt20.6%23.4%
DTBench89.1%91.2%
FrontierCode39.8%44.2%
FrontierMath Tier 461.0%39.0%
FrontierMath Tiers 1-382.1%72.2%
GPQA Diamond91.6%93.1%
LMCA48.5%52.7%
MirrorCode21.1%54.4%
OTIS Mock AIME98.3%97.2%
ProofBench60.0%87.0%
SciCode53.6%59.5%
SimpleBench46.8%60.7%
SimpleQA Verified41.0%50.6%
WebDev Arena15191674
WeirdML v35.2%7.5%

Specs and price

GPT-5.6 LunaKimi K3
Agent Index51.4 (#34)54.9 (#26)
Agent $/M$0.10$1.39
Input / output $/M$0.20 / $1.20$3.00 / $15.00
Context1.1M1M
ReleasedJul 9, 2026Jul 16, 2026
WeightsClosedOpen
API hosts529

Don’t pick one. Route.

Shogo can run GPT-5.6 Luna and Kimi K3 side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo