SShogo Bench

DeepSeek V4.1 Flash vs GPT-5.6 Luna

GPT-5.6 Luna leads DeepSeek V4.1 Flash by 0.2 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. DeepSeek V4.1 Flash wins 10 and GPT-5.6 Luna wins 8 of 19 shared benchmarks. GPT-5.6 Luna is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarDeepSeek V4.1 FlashGPT-5.6 Luna
Agentic coding5759
Long-horizon autonomy6261
Knowledge-work agents3939
Computer use4040
Security & context learning5151

Shared benchmarks

BenchmarkDeepSeek V4.1 FlashGPT-5.6 Luna
APEX-Agentsindex39.5%43.0%
ARC-AGI94.5%88.0%
ARC-AGI-272.9%59.5%
Chess Puzzles48.0%43.0%
CritPt14.3%20.6%
DTBench89.9%89.1%
FrontierCode43.9%39.8%
FrontierMath Tier 426.8%61.0%
FrontierMath Tiers 1-367.4%82.1%
GPQA Diamond89.8%91.6%
LMCA47.0%48.5%
MirrorCode64.4%21.1%
OTIS Mock AIME98.3%98.3%
ProofBench54.0%60.0%
SciCode51.9%53.6%
SimpleBench66.7%46.8%
SimpleQA Verified42.9%41.0%
WebDev Arena16141519
WeirdML v35.9%5.2%

Specs and price

DeepSeek V4.1 FlashGPT-5.6 Luna
Agent Index51.2 (#35)51.4 (#34)
Agent $/M$0.12$0.10
Input / output $/M$0.30 / $1.20$0.20 / $1.20
Context1M1.1M
ReleasedSep 10, 2026Jul 9, 2026
WeightsOpenClosed
API hosts395

Don’t pick one. Route.

Shogo can run DeepSeek V4.1 Flash and GPT-5.6 Luna side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo