Grok 4.20

ModelActive

by xAI · family “grok”

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... (description from the OpenRouter listing)

in: textin: imagein: fileout: textReasoningTool useStructured output

Epoch Capabilities Index

152.0 #40 of 268

90% CI 149.2 – 154.5

Listing ↗

Released

17 Feb 2026

Context window

2M tokens

Max output

1.8M tokens

Input $ / 1M tokens

$1.25

Output $ / 1M tokens

$2.50

Cached input $ / 1M

$0.20

Each benchmark shown separately with its own source

Benchmark results (15)

All benchmarks

Benchmark	Domain	Score	vs best recorded	Setting	Run	Source
LMCA	agents	38.7%	59%	—	—	External ↗
DTBench	reasoning	90.1%	92%	—	—	External ↗
FrontierMath-Tiers-1-3-v2-Private	math	44.9% ±3.0	48%	—	13 Jul 2026	Eval log ↗
FrontierMath-Tier-4-v2-Private	math	17.1% ±5.9	17%	—	13 Jul 2026	Eval log ↗
CL-bench Life	long-context	11.9%	54%	—	—	External ↗
CL-bench	long-context	22.2%	80%	—	—	External ↗
ProofBench	math	14.0%	14%	—	—	External ↗
Chess Puzzles	games	24.0% ±4.3	33%	—	13 Jul 2026	Epoch ↗
SimpleQA Verified	knowledge	30.2% ±1.5	40%	—	27 Aug 2026	Eval log ↗
Terminal Bench	agents	57.3%	68%	agent: Grok CLI	10 Mar 2026	External ↗
ARC-AGI-2	reasoning	65.1%	69%	—	—	External ↗
WeirdML	coding	52.3% ±0.0	56%	—	—	External ↗
OTIS Mock AIME 2024-2025	math	92.2% ±3.3	92%	—	13 Jul 2026	Epoch ↗
GPQA diamond	science	89.3% ±1.9	93%	—	13 Jul 2026	Epoch ↗
ARC-AGI	reasoning	89.5%	91%	—	—	External ↗

Source: Epoch AI Benchmarking Hub (CC BY 4.0). “External” rows are leaderboard results Epoch collects from third parties. Best reported setting per benchmark is shown.

API list price over time

Price history

$ per 1M tokens

7 observations (OpenRouter listing + Internet Archive snapshots). History accumulates with every ingest run; a single point means no change has been observed yet.