AI Model Release Tracker

Grok 4.20 Beta

Released

Grok 4.20 Beta is an AI model released by SpaceXSI on Tuesday, Feb 17 2026, 8 days after Composer 1.5. Benchmark results (shown below) cover BullshitBench v2, SWE-Bench Verified, and ARC-AGI-2.

API pricing

Under 200K input tokens
InputWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it.
$1.25
Cached inputA reduced rate for text you send over and over. If every request starts with the same instructions or the same document, the provider keeps a copy ready and charges less to read it again.
$0.20
OutputWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one.
$2.50
At least 200K input tokens
InputWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it.
$2.50
Cached inputA reduced rate for text you send over and over. If every request starts with the same instructions or the same document, the provider keeps a copy ready and charges less to read it again.
$0.40
OutputWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one.
$5.00
  • Reasoning or thinking is supported.
USD per 1M tokensEvery price here is for one million tokens. A token is roughly three-quarters of a word, so a million tokens is about 750,000 words of text.Model mappingVerified October 19, 2026Read from docs.x.ai. Click the check to open the source.

Benchmarks

Coding

SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better.via BenchLM
76.7%
#25 of 70

Reasoning & science

ARC-AGI-2Abstract reasoning — Puzzle-style tests of abstract reasoning and pattern-finding — the kind of thing people find easy but AIs often struggle with. Higher is better.via BenchLM
53.3%
#32 of 41

Robustness

BullshitBench v2Nonsense detection — Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better.
56%
#32 of 98

Source: BenchLM, retrieved 31 August 2026. Other sources are identified on the linked benchmark pages.

Compare Grok 4.20 Beta with

Grok 4.20 Beta

Suggested comparisons

Grok 4.20 BetavsGrok 4.7Grok 4.20 BetavsGPT-6.1 TerraGrok 4.20 BetavsClaude Haiku 5.5Grok 4.20 BetavsGemini 4 CarbonGrok 4.20 BetavsMuse Spark OpenGrok 4.20 BetavsDeepSeek-V4.1-Flash-1017Grok 4.20 BetavsMistral Large 4Grok 4.20 BetavsKimi K3.1 ProGrok 4.20 BetavsGLM-5.3-FlashGrok 4.20 BetavsQwen3.8-Max-0902Grok 4.20 BetavsNemotron 3.5 Bolt E4B 4GB

About

Grok 4.20 Beta marks an early release stage of xAI's 4.20 model line. The later system card describes both single-agent and multi-agent deployment, along with text and image inputs and text output. Its intended uses included research, writing, coding and reasoning, extending Grok's general assistant role rather than defining a model limited to one professional domain.

The beta name should be read with the checkpoint and product stage in mind. Subsequent API documentation lists dated production identifiers and aliases for the 4.20 family, which do not establish that every later specification was available in the initial beta. Likewise, generating images through Grok Imagine was a tool integration in the hosted product, distinct from the language model's own text output.

Frequently asked questions

Grok 4.20 Beta was released by SpaceXSI on Tuesday, Feb 17 2026.

All SpaceXSI releases

21 tracked

2023

1 release
Grok‑1
Nov 3 2023