BenchLeader

Grok 4.5 vs Muse Spark 1.3

Verdict
  • Grok 4.5 and Muse Spark 1.3 (xhigh) are level on quality (63.4 vs 63.9).
  • Grok 4.5 is stronger in human preference, knowledge, reasoning.
  • Muse Spark 1.3 (xhigh) is stronger in agents & tools, coding, composite, long context, multimodal.
  • Muse Spark 1.3 (xhigh) is 1.5× cheaper ($2.00 vs $3.00 per 1M blended).
  • Muse Spark 1.3 (xhigh) streams 3.5× faster (193 vs 56 tokens per second).
MetricGrok 4.5Muse Spark 1.3 (xhigh)
BenchLeader Index63.463.9
Agents & tools score58.660.4
Coding score59.970.8
Composite score66.078.6
Human preference score66.9
Knowledge score76.075.0
Long context score66.268.1
Multimodal score64.866.6
Reasoning score68.9
Blended price $/M$3.00$2.00
Output speed56 tok/s193 tok/s
Time to first answer12.7 s40.0 s
Context window500k1M
SimpleBench70.0%
WeirdML46.4%
APEX-Agents34.2%
FrontierCode42.4%
Epoch Capabilities Index153.9
LMArena Text1471
LMArena Hard Prompts1495
LMArena Coding1523
LMArena WebDev15561625
LMArena Vision1291
LMArena Agent3.9
LiveBench75.8%81.6%
LiveBench Reasoning87.2%89.7%
LiveBench Coding68.6%81.1%
LiveBench Agentic Coding56.5%64.1%
LiveBench Mathematics90.8%96.0%
LiveBench Data Analysis73.0%79.6%
LiveBench Language82.8%82.8%
AA Intelligence Index39.145.2
AA-LCR79.3%83.0%
MMMU-Pro80.4%82.0%
AA-Omniscience25.323.1
GPQA Diamond (AA)93.1%94.1%
Humanity's Last Exam (AA)42.7%47.5%
SciCode (AA)55.0%59.7%
Terminal-Bench 2.1 (Vals)72.3%
Vals Index60.3
Kagi LLM Benchmark83.5%

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