SShogo Bench

Kimi K3 vs Muse Spark 1.1

Kimi K3 leads Muse Spark 1.1 by 6.6 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Kimi K3 wins 8 and Muse Spark 1.1 wins 3 of 11 shared benchmarks. Muse Spark 1.1 is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarKimi K3Muse Spark 1.1
Agentic coding6353
Long-horizon autonomy6564
Knowledge-work agents4234
Computer use4336
Security & context learning5348

Shared benchmarks

BenchmarkKimi K3Muse Spark 1.1
APEX-Agentsindex47.8%31.8%
DeepSWEindex68.5%53.3%
Vending-Bench 2index$5,165$6,520
Chess Puzzles37.0%34.0%
CritPt23.4%15.1%
DTBench91.2%94.4%
LMCA52.7%49.9%
ProofBench87.0%39.0%
SciCode59.5%58.8%
SimpleQA Verified50.6%57.8%
WebDev Arena16741542

Specs and price

Kimi K3Muse Spark 1.1
Agent Index54.9 (#26)48.3 (#40)
Agent $/M$1.39$0.50
Input / output $/M$3.00 / $15.00$1.25 / $4.25
Context1M1M
ReleasedJul 16, 2026Apr 8, 2026
WeightsOpenClosed
API hosts291

Don’t pick one. Route.

Shogo can run Kimi K3 and Muse Spark 1.1 side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo