Models, measured side by side
Grok 4.20 vs Llama 4 Scout
Compare Grok 4.20, Llama 4 Scout by token pricing, context length, benchmark results, speed and tool support.
Monthly workload
Monthly cost breakdown
InputOutput
Grok 4.20
$38 / mo
Llama 4 Scout
$3.50 / mo
Performance comparison
Grok 4.20
25.7
Llama 4 Scout
8.1
Overview
Pricing & capacity
| Pricing & capacity | Grok 4.20 | Llama 4 Scout |
|---|---|---|
| Input / 1M tokens | $1.25 | $0.10 |
| Output / 1M tokens | $2.50 | $0.30 |
| Cached input / 1M | $0.20 | Not listed |
| Monthly cost | $38 | $3.50 |
| Context window | 2M | 1.3M |
| Maximum output | 1.8M | 16K |
Input / 1M tokens
- Grok 4.20
- $1.25
- Llama 4 Scout
- $0.10
Output / 1M tokens
- Grok 4.20
- $2.50
- Llama 4 Scout
- $0.30
Cached input / 1M
- Grok 4.20
- $0.20
- Llama 4 Scout
- Not listed
Monthly cost
- Grok 4.20
- $38
- Llama 4 Scout
- $3.50
Context window
- Grok 4.20
- 2M
- Llama 4 Scout
- 1.3M
Maximum output
- Grok 4.20
- 1.8M
- Llama 4 Scout
- 16K
Benchmarks & performance
| Benchmarks & performance | Grok 4.20 | Llama 4 Scout |
|---|---|---|
| Arena Elo | Not listed | Not listed |
| MMLU-Pro | Not listed | Not listed |
| Intelligence Index | 25.7 | 8.1 |
| GPQA | 91.1% | 58.7% |
| Humanity’s Last Exam | 34.5% | 3.8% |
| SciCode | Not listed | 21.3% |
| Output speed | 120 tokens/s | 69 tokens/s |
| Time to first token | 19.5 s | 0.86 s |
| DesignArena rating | Not listed | 793 |
| DesignArena win rate | Not listed | 26.6% |
Arena Elo
- Grok 4.20
- Not listed
- Llama 4 Scout
- Not listed
MMLU-Pro
- Grok 4.20
- Not listed
- Llama 4 Scout
- Not listed
Intelligence Index
- Grok 4.20
- 25.7
- Llama 4 Scout
- 8.1
GPQA
- Grok 4.20
- 91.1%
- Llama 4 Scout
- 58.7%
Humanity’s Last Exam
- Grok 4.20
- 34.5%
- Llama 4 Scout
- 3.8%
SciCode
- Grok 4.20
- Not listed
- Llama 4 Scout
- 21.3%
Output speed
- Grok 4.20
- 120 tokens/s
- Llama 4 Scout
- 69 tokens/s
Time to first token
- Grok 4.20
- 19.5 s
- Llama 4 Scout
- 0.86 s
DesignArena rating
- Grok 4.20
- Not listed
- Llama 4 Scout
- 793
DesignArena win rate
- Grok 4.20
- Not listed
- Llama 4 Scout
- 26.6%
Tools & features
| Tools & features | Grok 4.20 | Llama 4 Scout |
|---|---|---|
| Function calling | Supported | Supported |
| Structured outputs | Supported | Supported |
| JSON mode | Supported | Supported |
| Reasoning | Supported | Not listed |
| Built-in web search | Not listed | Not listed |
| Log probabilities | Supported | Not listed |
| Deterministic seed | Supported | Supported |
| Parallel tool calls | Not listed | Not listed |
| Prompt caching | Supported | Not listed |
Function calling
- Grok 4.20
- Supported
- Llama 4 Scout
- Supported
Structured outputs
- Grok 4.20
- Supported
- Llama 4 Scout
- Supported
JSON mode
- Grok 4.20
- Supported
- Llama 4 Scout
- Supported
Reasoning
- Grok 4.20
- Supported
- Llama 4 Scout
- Not listed
Built-in web search
- Grok 4.20
- Not listed
- Llama 4 Scout
- Not listed
Log probabilities
- Grok 4.20
- Supported
- Llama 4 Scout
- Not listed
Deterministic seed
- Grok 4.20
- Supported
- Llama 4 Scout
- Supported
Parallel tool calls
- Grok 4.20
- Not listed
- Llama 4 Scout
- Not listed
Prompt caching
- Grok 4.20
- Supported
- Llama 4 Scout
- Not listed
API & availability
| API & availability | Grok 4.20 | Llama 4 Scout |
|---|---|---|
| API identifier | grok-4.20 | llama-4-scout |
| API providers | Not listed | DeepInfra, Novita, Google |
| Knowledge cutoff | 2025-09-01 | 2024-08-31 |
| Prices checked | Oct 8, 2026 | Oct 8, 2026 |
API identifier
- Grok 4.20
grok-4.20- Llama 4 Scout
llama-4-scout
API providers
- Grok 4.20
- Not listed
- Llama 4 Scout
- DeepInfra, Novita, Google
Knowledge cutoff
- Grok 4.20
- 2025-09-01
- Llama 4 Scout
- 2024-08-31
Prices checked
- Grok 4.20
- Oct 8, 2026
- Llama 4 Scout
- Oct 8, 2026
About the models
Grok 4.20
Grok 4.20 is a reasoning model from SpaceXAI with industry-leading speed and agentic tool calling capabilities. It combines the lowest hallucination rate on the market with strict prompt adherance, delivering...
Full pricing & detailsLlama 4 Scout
Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B. It supports native multimodal input...
Full pricing & detailsComparison FAQ
Grok 4.20: $1.25 input and $2.50 output per million tokens. Llama 4 Scout: $0.10 input and $0.30 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.