BenchLeader

Kimi K3 vs Muse Spark

Verdict
  • Kimi K3 leads on quality: 66.5 vs 65.4.
  • Kimi K3 is stronger in agents & tools, composite, human preference, knowledge, long context, maths.
  • Muse Spark is stronger in coding, multimodal, instruction following, reasoning.
MetricKimi K3Muse Spark
BenchLeader Index66.565.4
Agents & tools score67.566.3
Coding score64.966.0
Composite score74.168.4
Human preference score71.269.0
Knowledge score68.164.5
Long context score71.165.5
Maths score78.156.3
Multimodal score65.165.9
Instruction following score79.6
Reasoning score70.0
Blended price $/M$6.00
Output speed37 tok/s
Time to first answer57.7 s
Context window1.0M262k
GPQA Diamond89.8%
OTIS Mock AIME88.9%
Humanity's Last Exam40.6%
SciCode58.7%51.5%
APEX-Agents39.3%
FrontierCode44.2%
ProofBench87.0%17.0%
Epoch Capabilities Index157.6152.1
LMArena Text1488
LMArena Hard Prompts1505
LMArena Coding1526
LMArena Vision1306
LiveBench79.2%
LiveBench Reasoning90.7%
LiveBench Coding81.5%
LiveBench Agentic Coding62.2%
LiveBench Mathematics84.4%
LiveBench Data Analysis78.7%
LiveBench Language85.5%
AA Intelligence Index43.831.3
IFBench75.9%
AA-LCR88.7%78.0%
MMMU-Pro80.5%80.5%
AA-Omniscience19.77.2
Terminal-Bench Hard45.5%
GPQA Diamond (AA)93.5%88.4%
Humanity's Last Exam (AA)46.9%40.7%
SciCode (AA)59.5%
τ²-Bench Telecom (AA)91.5%
AIME (Vals)96.9%
MMLU-Pro87.3%
LegalBench86.0%84.2%
CorpFin71.6%65.1%
TaxEval75.7%77.7%
Terminal-Bench 2.1 (Vals)80.9%
SWE-bench (Vals)93.4%74.4%
GPQA Diamond (Vals)92.9%89.7%
Vals Index57.8
SWE-Bench Pro55.0%
MCP Atlas82.2%
MultiChallenge75.5%
PRBench Finance52.4%
PRBench Legal52.3%
MultiNRC59.0%
TutorBench68.5%
EQ-Bench 41339

Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.