Pass IndexThe State of AISign in

Grok 4.20

xAI's Grok 4.20 reasoning variant (grok-4.20-0309-reasoning), described on docs.x.ai as "a high-performance model with industry-leading speed and agentic tool calling capabilities" combining "the lowest hallucination rate on the market with strict prompt adherence.".

text + image + file → text · made by xAI

$1.25$2.5per Mtok in / out
xAI · 8 sellers

Sold by 13 ways

SellerLaneRate
xAI apibatch$1per Mtok in$2per Mtok outdocs.x.ai · read 2026-08-25
Google Vertex AI 0 ways$1.25per Mtok in$2.5per Mtok out$0.2per Mtok cached
Google Vertex AI cloudnot on the seller's list todaystandard$1.25per Mtok in$2.5per Mtok out$0.2per Mtok cachedcloud.google.com · read 2026-08-24
xAI apistandard$1.25per Mtok in$2.5per Mtok out$0.2per Mtok cacheddocs.x.ai · read 2026-08-25
Microsoft Azure AI cloudstandard$2per Mtok in$6per Mtok out$1.25per Mtok cachedmodels.dev · read 2026-09-16
xAI apilong-context$2.5per Mtok in$5per Mtok out$0.4per Mtok cacheddocs.x.ai · read 2026-08-25
Oracle Cloud Infrastructure cloudstandard$3per Mtok in$15per Mtok outraw.githubusercontent.com · read 2026-09-16
Nous Research aggregatorstandard$1.25per Mtok in$2.5per Mtok out$0.2per Mtok cachedinference-api.nousresearch.com · read 2026-09-12
OpenRouter 3 ways$1.25per Mtok in$2.5per Mtok out$0.2per Mtok cached
OpenRouter aggregatorstandard$1.25per Mtok in$2.5per Mtok out$0.2per Mtok cachedopenrouter.ai · read 2026-08-24
OpenRouter aggregatorxai/priority$2.5per Mtok in$5per Mtok out$0.4per Mtok cachedopenrouter.ai · read 2026-08-25
OpenRouter aggregatorxai/zdr/priority$2.5per Mtok in$5per Mtok out$0.4per Mtok cachedopenrouter.ai · read 2026-08-25
Vercel AI Gateway aggregatorstandard$1.25per Mtok in$2.5per Mtok out$0.2per Mtok cachedai-gateway.vercel.sh · read 2026-08-25
Venice AI aggregatorstandard$1.42per Mtok in$2.83per Mtok out$0.23per Mtok cachedmodels.dev · read 2026-09-16
ElectronHub aggregatorstandard$4per Mtok in$8per Mtok outapi.electronhub.ai · read 2026-09-16

Measured 41 standings

PlaceBoardMetricScore
3rdof 23Cl bench — Epoch AIOverall0.222
9thof 81Forecastbench — Epoch AIOverall score61.4
12thof 17Cl bench life — Epoch AIOverall0.119
18thof 36Design Arena · htmlslides (agents)Elo1180
18thof 34LMArena · SearchScore (Elo)1189
22ndof 24Blueprint bench 2 — Epoch AIScore0
22ndof 28τ³-Banking LeaderboardPass^118
28thof 60Vending bench 2 — Epoch AIScore4662.847
30thof 402LMArena · TextScore (Elo)1475
32ndof 161Dtbench — Epoch AIAccuracy90.13
35thof 112Ale bench — Epoch AIPerformance1150.28
36thof 39ARC-AGI-3 (Semi-Private)Score (%)0.09
37thof 42Design Arena · godotgamedev (agents)Elo1089
39thof 123Lmca — Epoch AIScore38.66
41stof 62Design Arena · webapps (agents)Elo1159
42ndof 92Design Arena · asciiart (models)Elo1193
42ndof 152LMArena · VisionScore (Elo)1256
44thof 106Frontiermath tiers 1 3 v2 — Epoch AImean_score44.912
45thof 62Frontiermath tier 4 v2 — Epoch AImean_score17.073
46thof 313GPQA Diamondmean_score89.331
47thof 64Proofbench — Epoch AIAccuracy14
51stof 55Design Arena · androidnative (agents)Elo1059
51stof 64Design Arena · fullstack (agents)Elo1063
53rdof 109Design Arena · svg (models)Elo1189
54thof 290OTIS Mock AIME 2024-2025mean_score92.222
56thof 227Arc agi 2 — Epoch AIScore65.14
57thof 63Design Arena · mobileapps (agents)Elo1118
58thof 202Terminalbench — Epoch AIAccuracy mean0.573
60thof 147Design Arena · 3d (models)Elo1221
60thof 80Simpleqa verified — Epoch AImean_score30.2
61stof 156Design Arena · codecategories (models)Elo1236
61stof 162Design Arena · website (models)Elo1242
63rdof 152Design Arena · dataviz (models)Elo1234
65thof 244Arc agi — Epoch AIScore89.5
65thof 170Weirdml — Epoch AIAccuracy52.26
66thof 155Design Arena · gamedev (models)Elo1216
67thof 150Design Arena · uicomponent (models)Elo1213
68thof 222Chess puzzles — Epoch AImean_score24
86thof 121Webdev arena — Epoch AIArena Score1373.56
87thof 128LMArena · WebDevScore (Elo)1374
184thof 656Epoch capabilities index — Epoch AIECI Score152.01

About

Grok 4.20 — a text model from xAI, sold by 8 companies from $1.25 in and $2.5 out per million tokens, placed 30th of 402 on LMArena · Text.

It takes text, images and files and returns text, with a context window of 2,000,000 tokens. It was published in March 2026, trained on material up to September 2025. Its sellers say it can reason step by step and call a tool. The catalogue files it under reasoning, code and agents. Eight companies sell it. The cheapest is $1.25 in and $2.5 out per million tokens at Nous Research. Beside the standard rate there are faster, batch and separately routed lanes. It has been measured on 41 boards, and stands best at 30th of 402 on LMArena · Text.

Every current figure

Maker
xAI
Register
model
Takes
text + image + file
Returns
text
Context
2,000,000 tokens
Published
March 2026
Knowledge to
September 2025
Licence
not read
Sellers
8
Price
$1.25 in and $2.5 out per million tokens — Nous Research
Boards
41
Best place
30th of 402 — LMArena · Text

Known as 10 names

grok-4.20-beta1grok-4.20-beta-0309-reasoningGrok 4.2x-ai/grok-4.20spacexai/grok-4.20-reasoningGrok 4.20 Reasoninggrok-4.20-0309-reasoninggrok-4.20-reasoninggrok-4.20-reasoning-latestSpaceXAI: Grok 4.20