SShogo Bench

Gemini 3.8 Flash vs GPT-5.6 Sol

GPT-5.6 Sol leads Gemini 3.8 Flash by 5.6 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Gemini 3.8 Flash wins 7 and GPT-5.6 Sol wins 16 of 24 shared benchmarks. Gemini 3.8 Flash is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarGemini 3.8 FlashGPT-5.6 Sol
Agentic coding6371
Long-horizon autonomy6472
Knowledge-work agents4746
Computer use4453
Security & context learning5358

Shared benchmarks

BenchmarkGemini 3.8 FlashGPT-5.6 Sol
APEX-Agentsindex64.3%51.4%
DeepSWEindex73.8%72.7%
Vending-Bench 2index$5,094$9,619
FrontierSWEindex19.6%32.2%
CursorBenchindex39.6%41.7%
ARC-AGI98.5%97.5%
ARC-AGI-289.2%92.5%
Chess Puzzles58.0%68.0%
CritPt18.3%32.3%
DTBench95.7%96.0%
FrontierCode41.2%47.5%
FrontierMath Tier 422.0%82.9%
FrontierMath Tiers 1-368.4%89.1%
GeoBench86.0%73.0%
GPQA Diamond95.4%93.5%
LMCA52.9%59.2%
MirrorCode38.9%20.0%
OTIS Mock AIME98.9%100.0%
ProofBench48.0%83.0%
SciCode56.6%57.8%
SimpleBench82.4%71.7%
SimpleQA Verified69.7%69.7%
WebDev Arena15681617
WeirdML v37.4%15.0%

Specs and price

Gemini 3.8 FlashGPT-5.6 Sol
Agent Index56.2 (#21)61.8 (#14)
Agent $/M$0.35$1.85
Input / output $/M$0.75 / $3.75$4.00 / $20.00
Context1M1.1M
ReleasedSep 2, 2026Aug 7, 2026
WeightsClosedClosed
API hosts65

Don’t pick one. Route.

Shogo can run Gemini 3.8 Flash and GPT-5.6 Sol side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo