Models, measured side by side
Kimi K3 vs Muse Spark 1.3
Compare Kimi K3, Muse Spark 1.3 by token pricing, context length, benchmark results, speed and tool support.
Monthly workload
Monthly cost breakdown
InputOutput
Kimi K3
$122 / mo
Muse Spark 1.3
$46 / mo
Performance comparison
Kimi K3
43.6
Muse Spark 1.3
48.1
Overview
| Overview | Kimi K3 | Muse Spark 1.3 |
|---|---|---|
| Provider | MoonshotAI | Meta |
| Released | 2026-07-16 | 2026-09-02 |
| Status | Active | Active |
| Input formats | textimagevideo | textimagevideofile |
| Output formats | text | text |
Provider
- Kimi K3
- MoonshotAI
- Muse Spark 1.3
- Meta
Released
- Kimi K3
- 2026-07-16
- Muse Spark 1.3
- 2026-09-02
Status
- Kimi K3
- Active
- Muse Spark 1.3
- Active
Input formats
- Kimi K3
- textimagevideo
- Muse Spark 1.3
- textimagevideofile
Output formats
- Kimi K3
- text
- Muse Spark 1.3
- text
Pricing & capacity
| Pricing & capacity | Kimi K3 | Muse Spark 1.3 |
|---|---|---|
| Input / 1M tokens | $3.00 | $1.25 |
| Output / 1M tokens | $12 | $4.25 |
| Cached input / 1M | $0.43 | $0.15 |
| Monthly cost | $122 | $46 |
| Context window | 1M | 1M |
| Maximum output | 944K | 944K |
Input / 1M tokens
- Kimi K3
- $3.00
- Muse Spark 1.3
- $1.25
Output / 1M tokens
- Kimi K3
- $12
- Muse Spark 1.3
- $4.25
Cached input / 1M
- Kimi K3
- $0.43
- Muse Spark 1.3
- $0.15
Monthly cost
- Kimi K3
- $122
- Muse Spark 1.3
- $46
Context window
- Kimi K3
- 1M
- Muse Spark 1.3
- 1M
Maximum output
- Kimi K3
- 944K
- Muse Spark 1.3
- 944K
Benchmarks & performance
| Benchmarks & performance | Kimi K3 | Muse Spark 1.3 |
|---|---|---|
| Intelligence Index | 43.6 | 48.1 |
| GPQA | 93.5% | 93.5% |
| Humanity’s Last Exam | 46.9% | 48.7% |
| SciCode | 59.5% | 58.8% |
| Output speed | 45 tokens/s | 146 tokens/s |
| Time to first token | 3.83 s | 47.86 s |
| DesignArena rating | 1,368 | 1,358 |
| DesignArena win rate | 63.1% | 57.3% |
Intelligence Index
- Kimi K3
- 43.6
- Muse Spark 1.3
- 48.1
GPQA
- Kimi K3
- 93.5%
- Muse Spark 1.3
- 93.5%
Humanity’s Last Exam
- Kimi K3
- 46.9%
- Muse Spark 1.3
- 48.7%
SciCode
- Kimi K3
- 59.5%
- Muse Spark 1.3
- 58.8%
Output speed
- Kimi K3
- 45 tokens/s
- Muse Spark 1.3
- 146 tokens/s
Time to first token
- Kimi K3
- 3.83 s
- Muse Spark 1.3
- 47.86 s
DesignArena rating
- Kimi K3
- 1,368
- Muse Spark 1.3
- 1,358
DesignArena win rate
- Kimi K3
- 63.1%
- Muse Spark 1.3
- 57.3%
Tools & features
| Tools & features | Kimi K3 | Muse Spark 1.3 |
|---|---|---|
| Function calling | Supported | Supported |
| Structured outputs | Supported | Supported |
| JSON mode | Supported | Supported |
| Reasoning | Supported | Supported |
| Log probabilities | Supported | Not listed |
| Deterministic seed | Supported | Not listed |
| Prompt caching | Supported | Supported |
Function calling
- Kimi K3
- Supported
- Muse Spark 1.3
- Supported
Structured outputs
- Kimi K3
- Supported
- Muse Spark 1.3
- Supported
JSON mode
- Kimi K3
- Supported
- Muse Spark 1.3
- Supported
Reasoning
- Kimi K3
- Supported
- Muse Spark 1.3
- Supported
Log probabilities
- Kimi K3
- Supported
- Muse Spark 1.3
- Not listed
Deterministic seed
- Kimi K3
- Supported
- Muse Spark 1.3
- Not listed
Prompt caching
- Kimi K3
- Supported
- Muse Spark 1.3
- Supported
API & availability
| API & availability | Kimi K3 | Muse Spark 1.3 |
|---|---|---|
| API identifier | kimi-k3 | muse-spark-1.3 |
| Prices checked | Oct 8, 2026 | Oct 8, 2026 |
API identifier
- Kimi K3
kimi-k3- Muse Spark 1.3
muse-spark-1.3
Prices checked
- Kimi K3
- Oct 8, 2026
- Muse Spark 1.3
- Oct 8, 2026
About the models
Kimi K3
Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...
Full pricing & detailsMuse Spark 1.3
Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It is designed to keep track of information across extended tasks, work through...
Full pricing & detailsComparison FAQ
Kimi K3: $3.00 input and $12 output per million tokens. Muse Spark 1.3: $1.25 input and $4.25 output per million tokens. The cheaper choice depends on your input-to-output ratio; a model can have a lower input rate but a higher output rate.
