SShogo Bench

Grok 4.7 vs Kimi K3

Kimi K3 leads Grok 4.7 by 0.1 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Grok 4.7 wins 10 and Kimi K3 wins 9 of 20 shared benchmarks. Grok 4.7 is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarGrok 4.7Kimi K3
Agentic coding5863
Long-horizon autonomy6965
Knowledge-work agents4742
Computer use3943
Security & context learning5053

Shared benchmarks

BenchmarkGrok 4.7Kimi K3
APEX-Agentsindex70.0%47.8%
Vending-Bench 2index$10,537$5,165
FrontierSWEindex29.5%25.9%
ARC-AGI90.2%94.5%
ARC-AGI-261.4%60.4%
Chess Puzzles41.0%37.0%
CritPt18.0%23.4%
DTBench96.0%91.2%
FrontierCode47.6%44.2%
FrontierMath Tier 417.1%39.0%
FrontierMath Tiers 1-353.0%72.2%
GPQA Diamond92.7%93.1%
LMCA49.4%52.7%
MirrorCode54.4%54.4%
OTIS Mock AIME98.1%97.2%
ProofBench26.0%87.0%
SciCode57.8%59.5%
SimpleQA Verified56.0%50.6%
WebDev Arena16381674
WeirdML v38.5%7.5%

Specs and price

Grok 4.7Kimi K3
Agent Index54.8 (#27)54.9 (#26)
Agent $/M$0.98$1.39
Input / output $/M$2.00 / $6.00$3.00 / $15.00
Context500K1M
ReleasedSep 21, 2026Jul 16, 2026
WeightsClosedOpen
API hosts529

Don’t pick one. Route.

Shogo can run Grok 4.7 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