SShogo Bench

Claude Opus 5.5 vs Kimi K3

Claude Opus 5.5 leads Kimi K3 by 16.0 points on the Shogo Agent Index. Claude Opus 5.5 wins 22 and Kimi K3 wins 0 of 22 shared benchmarks. Kimi K3 is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarClaude Opus 5.5Kimi K3
Agentic coding8163
Long-horizon autonomy7665
Knowledge-work agents5842
Computer use6643
Security & context learning6253

Shared benchmarks

BenchmarkClaude Opus 5.5Kimi K3
APEX-Agentsindex73.5%47.8%
Vending-Bench 2index$9,235$5,165
FrontierSWEindex62.3%25.9%
PostTrainBenchindex43.8%31.5%
ARC-AGI98.5%94.5%
ARC-AGI-293.3%60.4%
Chess Puzzles67.0%37.0%
CritPt31.7%23.4%
DTBench98.9%91.2%
FrontierCode54.6%44.2%
FrontierMath Tier 495.0%39.0%
FrontierMath Tiers 1-391.2%72.2%
GPQA Diamond94.1%93.1%
LMCA68.2%52.7%
MirrorCode95.6%54.4%
OTIS Mock AIME100.0%97.2%
ProofBench100.0%87.0%
SciCode66.9%59.5%
SimpleBench88.4%60.7%
SimpleQA Verified72.2%50.6%
WebDev Arena18201674
WeirdML v331.2%7.5%

Specs and price

Claude Opus 5.5Kimi K3
Agent Index70.9 (#3)54.9 (#26)
Agent $/M$1.69$1.39
Input / output $/M$4.00 / $20.00$3.00 / $15.00
Context1M1M
ReleasedSep 22, 2026Jul 16, 2026
WeightsClosedOpen
API hosts629

Don’t pick one. Route.

Shogo can run Claude Opus 5.5 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