★ Reading this for free? Get 20 structured AI courses + per-chapter AI tutor — the first chapter of every course free, no card.Start free in 30 seconds

xAI · Closed-API Model

Grok 4.3 Review: xAI's Aggressive Price Cut, Tested

xAI launched Grok 4.3 on April 30, 2026 and the headline is the price: $1.25 input / $2.50 output per million tokens — roughly 4.8× cheaper than Grok 4's $3 / $15, and well under GPT-5.5 ($5 / $30). You get a 1M-token context window, always-on reasoning you cannot disable, and native video input. This review covers the pricing, benchmarks, what's actually new, and which open-weight models come closest if you need to self-host.

📅 Published: June 19, 2026🔄 Last Updated: June 19, 2026✓ Manually Reviewed

Note: Grok 4.3 is API-only — it cannot be downloaded or run locally. For self-hostable models with comparable performance, see DeepSeek V4, GLM-5, and Qwen3-Coder-Next.

Key takeaways

  • Price collapse: $1.25 / $2.50 per Mtok — ~4.8× cheaper than Grok 4, and cheaper than GPT-5.5 and Claude Opus 4.7.
  • 1M context, no fixed output cap — on par with Gemini 3.1 Pro and GPT-5.5.
  • Always-on reasoning — you cannot toggle it off; the model thinks before every reply (~100 tok/s).
  • Native video input — ingests mp4/mov/webm directly; plus a Custom Voices cloning suite.
  • Coding trails — roughly 14 points behind Claude Opus 4.7 on SWE-Bench Verified.

Quick verdict

Grok 4.3 is the cheapest frontier-tier closed model right now, and that — more than any benchmark — is why it matters. If you run high-volume agentic pipelines, process video, or just want a capable reasoning model without GPT-5.5-level token bills, Grok 4.3 is the value pick. Always-on reasoning and a 1M context make it a solid default for long-horizon tasks.

Where it loses: coding. On SWE-Bench Verified it trails Claude Opus 4.7 by roughly 14 points, so for production code the leader is elsewhere. And the cheap headline price only applies up to 200K tokens — long prompts cost more. For local-first deployment, DeepSeek V4-Pro gets you most of Grok 4.3's general capability at zero per-token cost.

Specs at a glance

VendorxAI
Release dateApril 30, 2026 (beta Apr 17; Custom Voices May 2)
API price$1.25 input / $2.50 output per 1M tokens (≤200K)
Context window1,000,000 tokens
Max outputNo fixed cap published
ReasoningAlways-on (cannot be disabled)
ModalitiesText · Image · Video (input) · File gen (PDF/PPTX/XLSX) · Voice
LicenseProprietary (xAI Terms of Service)
Local self-hostable?No
API model IDgrok-4.3
Knowledge cutoffDecember 2025
AccessxAI API (console.x.ai) · Grok app · X Premium+ · OpenRouter · Vercel AI Gateway

The price collapse

The story of Grok 4.3 is cost. xAI cut input pricing by ~58% and output by ~83% versus Grok 4, landing at a price that undercuts every other frontier closed model. Here is how it stacks up per million tokens:

ModelInputOutput
Grok 4.3$1.25$2.50
Grok 4 (predecessor)$3.00$15.00
Gemini 3.1 Pro~$2.00~$12.00
GPT-5.5 (standard)$5.00$30.00
Claude Opus 4.7Higher tierHigher tier

The catch: the headline rate applies to requests up to 200K tokens. Prompts above that are billed at a higher rate, so whole-codebase analysis at the full 1M context costs more per token than $1.25 suggests. For sustained high volume, a self-hosted open-weight model still wins on total cost.

Sources: xAI / VentureBeat launch coverage (Apr 30, 2026), OpenRouter and Vercel AI Gateway model pages. Competitor prices are list rates and change frequently — verify before budgeting.

Benchmarks

Grok 4.3 is a strong reasoning and agentic model that gives up some ground on pure coding. Where independent trackers disagree on exact figures, we've flagged it — treat coding numbers as approximate.

BenchmarkGrok 4.3Notes
GPQA Diamond (science)~90%Frontier-tier reasoning.
τ²-Bench (agentic tool use)~98%A standout — strong tool/agent behavior.
SWE-Bench Verified (coding)~low-70s%~14 pts behind Claude Opus 4.7 (~87.6%). Figures vary by harness.
Vending-Bench (long-horizon agent)LeadsReported ~1.26× over Opus 4.7 on long-sequence sim.
Throughput~100 tok/sAmong the faster frontier-tier API models.

Sources: Artificial Analysis, vals.ai, and OpenRouter Grok 4.3 benchmark pages (2026); Anthropic Opus 4.7 announcement for the SWE-Bench comparison. Exact SWE-Bench Verified figures for Grok 4.3 vary across trackers (low-70s%); we report a conservative range.

Native video & voice

The other big change in Grok 4.3 is multimodality. The vision encoder ingests video directly — mp4, mov, and webm (reported up to ~5 minutes and 1080p) — and handles transcription, speaker segmentation, object tracking, and motion in a single pass, instead of chaining a separate transcriber, chunker, and language model. That collapses a common multi-tool video pipeline into one API call.

Video input

Direct mp4/mov/webm ingestion. Testers noted occasional hallucinated frame descriptions and intermittent latency — verify on critical tasks.

File generation

Generates PDF, PPTX, and XLSX outputs, plus tool use, vision input, and web search for agentic workflows.

Custom Voices (voice cloning)

Shipped May 2, 2026 alongside Grok 4.3 — clones a voice from ~1 minute of speech in under two minutes, free on the xAI console, sharing TTS/voice-agent APIs with 80+ preset voices.

When to pick Grok 4.3 vs alternatives

WorkloadBest pickWhy
Cost-sensitive agentic pipelinesGrok 4.3$1.25/$2.50 per Mtok + always-on reasoning + ~98% τ²-Bench.
Video understandingGrok 4.3Native video input in one pass — few closed models match it.
Production codingClaude Opus 4.7SWE-Bench Verified leader (~87.6%); ~14 pts ahead of Grok 4.3.
Math + ChatGPT ecosystemGPT-5.595.2% AIME; custom GPTs, plugins, mature function calling.
Cheapest long-context analysisGemini 3.1 Pro1M context; competitive per-token cost at long context.
Privacy-required workloadsDeepSeek V4Self-hostable, MIT licensed, 1M context.

Open-weight alternatives you can run locally

Grok 4.3 is API-only — no published weights, nothing to download. Its low price helps, but it's still a per-token bill on someone else's servers, with your data leaving your network. If you need a frontier-class model you can self-host for privacy, predictable cost, or offline use, these come closest:

ModelLicenseBest for
DeepSeek V4-ProMITClosest general-purpose match; 1M context
GLM-5MITReasoning on a smaller hardware footprint
Qwen3-Coder-NextApache 2.0Coding-heavy workflows (where Grok 4.3 lags)
Kimi K2.6Modified MIT1T MoE; strong agentic + coding

A note on Grok 5

To be clear: Grok 5 has not been released. As of June 2026, xAI confirmed in its January 2026 funding announcement that the next model — reportedly ~6 trillion parameters with a Mixture-of-Experts architecture — was still in training, and the originally floated Q1/Q2 2026 launch windows passed without a release. Grok 4.3 is xAI's current flagship. Ignore any "Grok 5 benchmarks" you see floating around; nothing official has shipped.

Frequently asked questions

Can I run Grok 4.3 locally?
No. Grok 4.3 is a proprietary xAI model with no published weights. You access it only through the xAI API (console.x.ai), the Grok app, X Premium+, and third-party gateways like OpenRouter and Vercel AI Gateway. It cannot be downloaded or self-hosted. For frontier-class models you can run on your own hardware, the closest open-weight alternatives are DeepSeek V4-Pro (MIT licensed, 1M context), GLM-5 (MIT), and the Qwen3-Coder family.
How much does the Grok 4.3 API cost?
As of June 2026, Grok 4.3 is $1.25 per million input tokens and $2.50 per million output tokens for requests up to 200K tokens (higher rates apply above that threshold). That is a large drop from Grok 4 at $3 / $15 — roughly 4.8× cheaper output. It undercuts most frontier closed models: GPT-5.5 standard is $5 / $30, Claude Opus 4.7 is more expensive still, and Gemini 3.1 Pro is around $2 / $12. The price collapse is the headline of this release.
What changed from Grok 4 to Grok 4.3?
Three things stand out. 1) Price: xAI cut API costs to $1.25 / $2.50 per Mtok from Grok 4's $3 / $15. 2) Reasoning is now always-on — unlike earlier Grok models you cannot toggle it off, the model "thinks" before every response. 3) Native video: the vision encoder ingests video directly (mp4/mov/webm) rather than relying on upstream frame extraction. Context stayed at 1M tokens and the knowledge cutoff moved to December 2025. xAI also shipped a Custom Voices voice-cloning suite alongside it.
Grok 4.3 vs GPT-5.5 vs Claude Opus 4.7: which is best?
Different leaders for different jobs. Grok 4.3 wins on price ($1.25/$2.50 vs GPT-5.5's $5/$30) and is strong on agentic/long-horizon tasks and factual accuracy. Claude Opus 4.7 leads coding — it holds the SWE-Bench Verified gold standard (~87.6%), where Grok 4.3 trails by roughly 14 points. GPT-5.5 leads math (95.2% AIME) and the ChatGPT ecosystem. Pick by workload: cost-sensitive agentic pipelines and video → Grok 4.3, hardest coding → Claude Opus 4.7, math + ChatGPT tooling → GPT-5.5.
Is Grok 4.3 good at coding?
It is capable but not the leader. On SWE-Bench Verified, third-party trackers put Grok 4.3 around the low-70s percent — solid, but roughly 14 points behind Claude Opus 4.7 (~87.6%), the current coding gold standard. For function-level generation, unit tests, and simple bug fixes the gap narrows to a few points. Where Grok 4.3 shines is long-horizon agentic work: it reportedly outperforms Opus 4.7 on long-sequence simulation tasks like Vending-Bench. For coding-heavy production work, a self-hostable alternative like Qwen3-Coder-Next or Claude Opus 4.7 is the stronger pick.
How big is Grok 4.3's context window?
1,000,000 tokens — about 2,000 pages of text — with no fixed output-token cap. That matches Gemini 3.1 Pro and GPT-5.5 (both 1M) and far exceeds the 200K of Claude's lineup. Note that the cheapest pricing tier applies to requests up to 200K tokens; prompts above that are billed at a higher rate, so whole-codebase or full-corpus prompts cost more per token than the headline figure suggests.
Does Grok 4.3 support video?
Yes. Grok 4.3 added native video input: the vision encoder ingests mp4/mov/webm directly (reported up to ~5 minutes and 1080p) and handles transcription, speaker segmentation, object tracking, and motion in one pass, rather than chaining a separate transcriber and frame extractor. It also generates files (PDF, PPTX, XLSX) and ships with the Custom Voices cloning suite. Independent testers noted occasional hallucinated frame descriptions, so verify on critical video tasks.
Is Grok 5 out yet?
No. As of June 2026, Grok 5 has not shipped. xAI confirmed in its January 2026 funding announcement that the ~6-trillion-parameter Grok 5 was still in training, and the originally floated Q1/Q2 2026 windows passed without a release. Treat any "Grok 5" claims as unconfirmed — Grok 4.3 is xAI's current flagship.
When should I use an open-weight model instead of Grok 4.3?
Use an open-weight alternative when: 1) Your data cannot leave your network (regulated industries, IP-sensitive work). 2) You want predictable costs — one-time hardware replaces an ongoing API bill, even a cheap one. 3) You need offline operation or to avoid a single-vendor dependency. The closest self-hostable matches are DeepSeek V4-Pro for general work, Qwen3-Coder-Next for coding, and GLM-5 for reasoning on a smaller hardware footprint. Many teams keep Grok 4.3 for video/agentic edge cases and run local models for the routine bulk.

Cut your AI API bill — run open weights locally

Grok 4.3 is cheap, but it's still a metered API on someone else's servers. The Local AI Master deployment course shows you how to run open-weight alternatives like DeepSeek V4 and Qwen3-Coder-Next on your own hardware — unlimited inference, full data privacy, zero per-token cost.

See the deployment course →

Related models

🎯
AI Learning Path

Go from reading about AI to building with AI

20 structured courses. Hands-on projects. Runs on your machine. Start free.

Or own it for life — Lifetime $149 $599, pay once
LM

Written by the Local AI Master Team

The team behind Local AI Master

We build Local AI Master around practical, testable local AI workflows: model selection, hardware planning, RAG systems, agents, and MLOps. The goal is to turn scattered tutorials into a structured learning path you can follow on your own hardware.

✓ Local AI Curriculum✓ Hands-On Projects✓ Open Source Contributor
More on AI Models Directory
See the full AI Models Directory guide.
📚
Free · no account required

Grab the AI Starter Kit — career roadmap, cheat sheet, setup guide

No spam. Unsubscribe with one click.

🎯
AI Learning Path

Found your model? Now build something with it.

20 hands-on courses — RAG, agents, fine-tuning — all running locally. First chapter free, no card.

Or own it for life — Lifetime $149 $599, pay once
Free Tools & Calculators