Skip to main content

Model comparison

Kimi K2.7 Code vs Muse Spark

Data verified

Head-to-head evidence from 15 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

Moonshot AI
54.03/100
Margin
16.3pts
winning →
70.35/100
0 category wins0 category wins

Public leaderboard positions: Kimi K2.7 Code #92 (Estimated); Muse Spark #16 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Kimi K2.7 Code and Muse Spark share 15 comparable benchmark results. 0 of 8 categories are comparable. 8 results are unique to Kimi K2.7 Code; 24 to Muse Spark.

Updated July 27, 2026
Shared results
15
Kimi K2.7 Code only
8
Muse Spark only
24
Comparable categories
0 / 8

Benchmark data for Kimi K2.7 Code and Muse Spark is coming soon on BenchLM.

Confidence note. This is a partial-evidence comparison with 15 shared benchmark results across 5 evidence categories; 0 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

BenchLM has partial data for these models, but not enough overlapping benchmark coverage to produce a fair score-level comparison yet.

Muse Spark has the larger context window at 262K, compared with 256K for Kimi K2.7 Code.

Category breakdown

Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.

Category scores and score margins for Kimi K2.7 Code and Muse Spark
CategoryKimi K2.7 CodeΔMuse Spark
AgenticKimi K2.7 CodeNot measuredMarginNo overlapMuse Spark59.0
CodingKimi K2.7 CodeNot measuredMarginNo overlapMuse Spark67.8
ReasoningKimi K2.7 CodeNot measuredMarginNo overlapMuse Spark42.5
KnowledgeKimi K2.7 CodeNot measuredMarginNo overlapMuse Spark50.4
MathKimi K2.7 CodeNot measuredMarginNo overlapMuse Spark32.9
MultimodalKimi K2.7 CodeNot measuredMarginNo overlapMuse Spark82.5

Operational comparison

Runtime and commercial metrics are compared only when both models have a complete sourced value.

MetricKimi K2.7 CodeMuse SparkComparison
Input / output priceUSD per 1M tokensKimi K2.7 Code$0.95 input / $4 outputMuse SparkNot availableA complete price comparison is not available.
Generation speedtokens per secondKimi K2.7 CodeNot availableMuse SparkNot availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenKimi K2.7 CodeNot availableMuse SparkNot availableA complete latency comparison is not available.
Context windowmaximum listed tokensKimi K2.7 Code256KMuse Spark262KMuse Spark lists the larger context window.

Benchmark Deep Dive

Agentic
BenchmarkKimi K2.7 CodeMuse SparkResult
Kimi Claw 24/7Source 46.9%Not comparable
MCP AtlasSource 76%Not comparable
MCP Mark VerifiedSource 81.1%Not comparable
AA Agentic IndexSource 29.6%28.7%Kimi K2.7 Code leads
τ²-bench resultsSource 90.1%91.5%Muse Spark leads
GDPval-AASource 34.3%32.2%Kimi K2.7 Code leads
GDPval-AASource 11861143Kimi K2.7 Code leads
Terminal-Bench 2.0Source 59%Not comparable
DeepSearchQASource 74.8%Not comparable
CyberGymSource 43.5%Not comparable
Claw-EvalSource 63.8%Not comparable
Coding
BenchmarkKimi K2.7 CodeMuse SparkResult
Kimi Code Bench v2Source 62.0%Not comparable
ProgramBenchSource 53.6%Not comparable
MLS-Bench LiteSource 35.1%Not comparable
cursorBench32Source 49.7%Not comparable
AA Coding IndexSource 60.8%58.6%Kimi K2.7 Code leads
AA-SciCodeSource 47.5%51.5%Muse Spark leads
SWE-bench VerifiedSource 77.4%Not comparable
SWE-bench ProSource 52.4%Not comparable
LiveCodeBench ProSource 80.0%Not comparable
Vibe Code BenchSource 19.67%Not comparable
Reasoning
BenchmarkKimi K2.7 CodeMuse SparkResult
AA-LCRSource 66.3%69.7%Muse Spark leads
CritPtSource 10.0%11.3%Muse Spark leads
ARC-AGI-2Source 42.5%Not comparable
Knowledge
BenchmarkKimi K2.7 CodeMuse SparkResult
Artificial Analysis Intelligence IndexSource 42.0%43.1%Muse Spark leads
AA-GPQA DiamondSource 89.6%88.4%Kimi K2.7 Code leads
AA-HLESource 32.8%39.9%Muse Spark leads
AA-Omniscience IndexSource -10.7%4.1%Muse Spark leads
AA-Omniscience AccuracySource 38.6%44.6%Muse Spark leads
AA-Omniscience Hallucination RateSource 80.3%73.2%Muse Spark leads
GPQA-DSource 89.5%Not comparable
HLESource 50.4%Not comparable
HLE w/o toolsSource 42.8%Not comparable
HealthBench HardSource 42.8%Not comparable
MedXpertQA (Text)Source 52.6%Not comparable
Math
BenchmarkKimi K2.7 CodeMuse SparkResult
FrontierMath v2 (Tiers 1-3)Source 39.000%Not comparable
FrontierMath v2 (Tier 4)Source 14.600%Not comparable
Multimodal
BenchmarkKimi K2.7 CodeMuse SparkResult
Design Arena WebsiteSource 1300Not comparable
CharXivSource 86.4%Not comparable
MMMU-ProSource 80.4%Not comparable
ERQASource 64.7%Not comparable
SimpleVQASource 71.3%Not comparable
ScreenSpot ProSource 84.1%Not comparable
ZeroBenchSource 33.0%Not comparable
MedXpertQA (MM)Source 78.4%Not comparable
AA-MMMU-ProSource 80.5%Not comparable
Inst. Following
BenchmarkKimi K2.7 CodeMuse SparkResult
AA-IFBenchSource 63.1%75.9%Muse Spark leads
Frequently Asked Questions (3)

Can I compare Kimi K2.7 Code and Muse Spark on BenchLM yet?

Not fully yet. BenchLM is tracking both models, but the sourced benchmark breakdown for this comparison is still coming soon.

Why does this comparison show “coming soon”?

BenchLM only shows category winners and benchmark-level calls when we have sourced results that can be compared fairly. For these models, the public benchmark coverage is not complete enough yet.

What data is available for Kimi K2.7 Code and Muse Spark today?

Kimi K2.7 Code: $0.95 input / $4.00 output per 1M tokens Both model pages still include creator, context window, reasoning mode, and other metadata while benchmark coverage fills in.

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

Kimi K2.7 Code
API / mo$3,713
Self-host / mo$18,221
Break-even326M/day
Muse Spark
API / mo$0
Self-host / moNot listed
Break-even
Proprietary model — self-hosting not applicable.
Model the full break-even

Related Comparisons

Last updated: July 27, 2026

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

One email each week. Unsubscribe anytime.