Grok 4.20 Multi-Agent
xAI · released Mar 31, 2026
- Type
- Closed
- Input
- $1.25
- Output
- $2.50
- Cached
- None held
List price · per 1M tokens · xAI at 1M context · machine-readable source ↗
Our take
Written Sep 4, 2026Grok 4.20 Multi-Agent is xAI's flagship text model with a two-million-token request limit and built-in multi-agent tooling. It is a premium pick for coding and long-document work, though its creative and instruction-following scores trail its stronger suits.
Choose this for coding assistance where measured preference peaks, or for analysing documents up to two million tokens in a single request. Use the standard tier for throughput; the premium tier costs double for slower speed. Skip it if you need open weights, if creative writing or precise instruction-following are central, or if you want measured quality data beyond Arena scores.
The case for it
- Two-million-token request limit — ten times the 200,000 common among frontier models we list.
- Arena-measured coding preference leads its own skill profile by 37.7 points over its overall text score.
- Standard tier pricing is consistent across OpenRouter and xAI's own offer.
The case against it
- Creative writing and instruction following lag its stronger suits by more than 20 points each.
- Premium tier charges double for 19.5% slower throughput — a poor speed-for-price trade.
- No MMLU, GPQA or SWE-bench scores in our data; quality picture is Arena-only.
How good is it?
A closed text model from xAI for everyday questions, drafting and coding, with no weak spots flagged on the board.
- answering everyday questionsArena Text (overall) · 29th of 168
- drafting and editing proseArena Creative Writing · 31st of 168
- writing and completing codeArena Coding · 39th of 168
EverydayGeneral questions and everyday reasoning
Arena Text (overall)29th of 168 · 1470
CodingWriting and fixing code on its own
Arena Coding39th of 168 · 1508
Arena Coding is the only board that has scored it for this.
AgenticPlanning, calling tools, staying on task
Not yet scored on Arena Agent.
WritingDrafting and rewriting prose
Arena Creative Writing31st of 168 · 1447
Arena Creative Writing is the only board that has scored it for this.
These tests check whether a model follows instructions — a precondition for all the work above, but not a measure of how well that work is done.
Every published score for this model6 scoresEvery figure we hold, from 6 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Where to rent it
Prices checked between 1 hour and 35 days ago — each listing carries its own date.
xAI, through OpenRouter
Why this differs from the header. The strip above quotes xAI's own list price; this is xAI through OpenRouter, which lists it at a lower price.
- per 1M tokens
- $1.25 in / $2.50 out
- Context served
- 2M
- Throughput
- ~349 tok/s
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouterOpenRouter's own listing | $1.25 / $2.50checked 1 hour ago | 2M | not measured | Unknown | Unknown | Unknown |
| xAIDirect | $1.25 / $2.50checked 35 days ago | 1M1M max reply | not measured | Unknown | Unknown | Unknown |
| xAIThrough OpenRouter | $1.25 / $2.50checked 1 hour ago | 2M1.8M max reply | 349 tok/s | No | Yes30 days | Confirmed |
| xAIpriority tierThrough OpenRouter | $2.50 / $5.00checked 1 hour ago | 2M1.8M max reply | 333 tok/s | No | Yes30 days | Confirmed |
OpenRouterAcross the 4 listings we hold: 2 say they do not train on prompts, 0 say they do and 2 do not say. 2 appear in the zero-retention registry we check; the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouterOpenRouter's own listing | ✗ | ✓ | ✓ |
| xAIDirect | |||
| xAIThrough OpenRouter | ✗ | ✓ | ✓ |
| xAIpriorityThrough OpenRouter | ✗ | ✓ | ✓ |
Tool calling: 0 of 4 listings say yes, 3 say no, 1 publishes no parameter list. JSON output: 3 of 4 listings say yes, 1 publishes no parameter list. Strict schema: 3 of 4 listings say yes, 1 publishes no parameter list.
When we formed this view
Recent changes
What moved
input −38% ($2.00 → $1.25 per 1M tokens), output −58% ($6.00 → $2.50 per 1M tokens)What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- 1 of 4 listings publishes no parameter list, so what its API accepts is unknown to us.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 2 of 4 listings do not say whether they train on prompts.
- We hold no batch or off-peak rate for any of its listings.
Licence and identifiers
What the licence allowsWe hold no licence record for this model. Inside are the identifiers you need to pull it — its Hugging Face repo where we have one, our slug and a machine-readable card.
Licence
We hold no licence record for this model, and no record of published weights either — so we can neither summarise its terms nor point you at the weights.
Identifiers
- Takes in, gives back
- Text, images and documents in, text out
- Catalogue slug
- x-ai-grok-4-20-multi-agent