Grok 4.20 vs Grok 4.20 Multi-Agent
Grok 4.20 (SpaceXSI) and Grok 4.20 Multi-Agent (SpaceXSI) have the same list price: $1.25 per 1M input tokens and $2.50 per 1M output tokens. Both hold 2,000,000 tokens of context. Only Grok 4.20 supports tool calling. Both accept text, images and files and produce text.
SpaceXSI: Grok 4.20 and SpaceXSI: Grok 4.20 Multi-Agent both accept text, images, and files, produce text, and list 2,000,000-token contexts. Grok 4.20 lists tool support; Multi-Agent is described as designed for collaborative, agent-based workflows.
Compared from catalog data checked
Side by side
0.00 the better figure in a row| Specification | Grok 4.20 | Grok 4.20 Multi-Agent |
|---|---|---|
| Maker | SpaceXSI | SpaceXSI |
| Added to catalog | 31 Mar 2026 | 31 Mar 2026 |
| Input, per 1M tokens | $1.25 | $1.25 |
| Output, per 1M tokens | $2.50 | $2.50 |
| Input above 200K tokensPer 1M; long prompts | $2.50 | $2.50 |
| Output above 200K tokens | $5.00 | $5.00 |
| Cached input read, per 1M | $0.20 | $0.20 |
| Context windowTokens | 2,000,000 | 2,000,000 |
| Max outputTokens | 1,800,000 | 1,800,000 |
| Accepts | file, image, text | file, image, text |
| Produces | text | text |
| Tool calling | Yes | No |
| Structured outputs | Yes | Yes |
| Reasoning effort control | No | Yes |
List prices through OpenRouter, in US dollars. Each model's page has its variants, price history and sample outputs.
What the same task costs
List price for one request, from the token counts shown. The same four tasks are priced on every model page.
| Task | Grok 4.20 | Grok 4.20 Multi-Agent |
|---|---|---|
| Answer a short question500 in / 300 out | $0.0014 | $0.0014 |
| Summarise a 50-page PDF25,000 in / 800 out | $0.033 | $0.033 |
| Review a 2,000-line source file30,000 in / 2,000 out | $0.043 | $0.043 |
| Read a 300-page contract set150,000 in / 1,500 out | $0.191 | $0.191 |
Which to choose
Choose Grok 4.20 Multi-Agent if
- You want to set how much it reasons before it answers
Neither maker's quality claims are repeated here. The reliable test is your own task: ask both the same question and read the two answers.