Models, measured side by side
Grok 4.7 vs Kimi K3
Compare Grok 4.7, Kimi K3 by token pricing, context length, benchmark results, speed and tool support.
Monthly workload
Monthly cost breakdown
InputOutput
Grok 4.7
$70 / mo
Kimi K3
$122 / mo
Performance comparison
Grok 4.7
46.4
Kimi K3
43.6
Overview
| Overview | Grok 4.7 | Kimi K3 |
|---|---|---|
| Provider | xAI | MoonshotAI |
| Released | 2026-09-21 | 2026-07-16 |
| Status | Active | Active |
| Input formats | textimagefile | textimagevideo |
| Output formats | text | text |
Provider
- Grok 4.7
- xAI
- Kimi K3
- MoonshotAI
Released
- Grok 4.7
- 2026-09-21
- Kimi K3
- 2026-07-16
Status
- Grok 4.7
- Active
- Kimi K3
- Active
Input formats
- Grok 4.7
- textimagefile
- Kimi K3
- textimagevideo
Output formats
- Grok 4.7
- text
- Kimi K3
- text
Pricing & capacity
| Pricing & capacity | Grok 4.7 | Kimi K3 |
|---|---|---|
| Input / 1M tokens | $2.00 | $3.00 |
| Output / 1M tokens | $6.00 | $12 |
| Cached input / 1M | $0.50 | $0.43 |
| Monthly cost | $70 | $122 |
| Context window | 500K | 1M |
| Maximum output | 450K | 944K |
Input / 1M tokens
- Grok 4.7
- $2.00
- Kimi K3
- $3.00
Output / 1M tokens
- Grok 4.7
- $6.00
- Kimi K3
- $12
Cached input / 1M
- Grok 4.7
- $0.50
- Kimi K3
- $0.43
Monthly cost
- Grok 4.7
- $70
- Kimi K3
- $122
Context window
- Grok 4.7
- 500K
- Kimi K3
- 1M
Maximum output
- Grok 4.7
- 450K
- Kimi K3
- 944K
Benchmarks & performance
| Benchmarks & performance | Grok 4.7 | Kimi K3 |
|---|---|---|
| Intelligence Index | 46.4 | 43.6 |
| GPQA | Not listed | 93.5% |
| Humanity’s Last Exam | 43.1% | 46.9% |
| SciCode | 57.4% | 59.5% |
| Output speed | 83 tokens/s | 45 tokens/s |
| Time to first token | 43.17 s | 3.83 s |
| DesignArena rating | 1,269 | 1,368 |
| DesignArena win rate | 47.6% | 63.1% |
Intelligence Index
- Grok 4.7
- 46.4
- Kimi K3
- 43.6
GPQA
- Grok 4.7
- Not listed
- Kimi K3
- 93.5%
Humanity’s Last Exam
- Grok 4.7
- 43.1%
- Kimi K3
- 46.9%
SciCode
- Grok 4.7
- 57.4%
- Kimi K3
- 59.5%
Output speed
- Grok 4.7
- 83 tokens/s
- Kimi K3
- 45 tokens/s
Time to first token
- Grok 4.7
- 43.17 s
- Kimi K3
- 3.83 s
DesignArena rating
- Grok 4.7
- 1,269
- Kimi K3
- 1,368
DesignArena win rate
- Grok 4.7
- 47.6%
- Kimi K3
- 63.1%
Tools & features
| Tools & features | Grok 4.7 | Kimi K3 |
|---|---|---|
| Function calling | Supported | Supported |
| Structured outputs | Supported | Supported |
| JSON mode | Supported | Supported |
| Reasoning | Supported | Supported |
| Log probabilities | Supported | Supported |
| Deterministic seed | Supported | Supported |
| Prompt caching | Supported | Supported |
Function calling
- Grok 4.7
- Supported
- Kimi K3
- Supported
Structured outputs
- Grok 4.7
- Supported
- Kimi K3
- Supported
JSON mode
- Grok 4.7
- Supported
- Kimi K3
- Supported
Reasoning
- Grok 4.7
- Supported
- Kimi K3
- Supported
Log probabilities
- Grok 4.7
- Supported
- Kimi K3
- Supported
Deterministic seed
- Grok 4.7
- Supported
- Kimi K3
- Supported
Prompt caching
- Grok 4.7
- Supported
- Kimi K3
- Supported
API & availability
| API & availability | Grok 4.7 | Kimi K3 |
|---|---|---|
| API identifier | grok-4.7 | kimi-k3 |
| API providers | xAI, xAI, xAI, xAI, xAI | Not listed |
| Prices checked | Oct 8, 2026 | Oct 8, 2026 |
API identifier
- Grok 4.7
grok-4.7- Kimi K3
kimi-k3
API providers
- Grok 4.7
- xAI, xAI, xAI, xAI, xAI
- Kimi K3
- Not listed
Prices checked
- Grok 4.7
- Oct 8, 2026
- Kimi K3
- Oct 8, 2026
About the models
Grok 4.7
Grok 4.7 is SpaceXAI's flagship model for coding, agentic tasks, and knowledge work, succeeding Grok 4.6. It is particularly strong at long-running software engineering tasks, verifying its own work, and...
Full pricing & detailsKimi 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 & detailsComparison FAQ
Grok 4.7: $2.00 input and $6.00 output per million tokens. Kimi K3: $3.00 input and $12 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.