VC
Value Add VC
โšกHomePulseโšกHelpful Apps๐Ÿ“Blog
โ† Value Add PulseAI

xAI's Grok 4.3 Goes Generally Available on Amazon Bedrock -- the Cheapest US Frontier Model There

Amazon made xAI's Grok 4.3 generally available on Bedrock, making xAI the third major independent lab on the platform alongside Anthropic and OpenAI. Priced at $1.25/$2.50 per million input/output tokens with a 1-million-token context window, Grok 4.3 lands as the cheapest US-lab frontier reasoning model on Bedrock.

Grok 4.3
Model
Amazon Bedrock (GA)
Platform
$1.25 / $2.50 per M tokens
Pricing
1M tokens
Context
Mantle engine
Inference
TC
Trace Cohen
Early-stage VC & angel ยท Founder, New York Venture Partners
June 21, 2026
1 min read
ShareXLinkedInEmail
THE RUNDOWN
1

Enterprises can now reach Grok through AWS's procurement and security stack -- a huge distribution unlock for xAI

2

Aggressive pricing pressures Anthropic and OpenAI on the cost-per-token battleground

3

Bedrock becoming a multi-lab marketplace turns model choice into a commodity selection for enterprises

4

It runs on Amazon's new Mantle inference engine, signaling AWS's push on price-performance

TC
The VC Read ยท Trace's TakeTrace Cohen

The model layer is commoditizing in real time, and Bedrock is where you can watch it happen -- three frontier labs on one shelf, competing on price, latency and trust. Grok undercutting Anthropic and OpenAI on cost-per-token is the tell: when capability converges, distribution and price win, and AWS owns the distribution. For founders, the lesson is that betting your product on a single model API is a strategic risk -- the abstraction layer that lets you swap models is becoming essential. xAI getting enterprise reach through AWS is also how Grok escapes the 'Musk's chatbot' framing into serious B2B.

๐Ÿค– AI Landscape โ†’๐Ÿ“ˆ AI Valuations โ†’

Amazon has made xAI's Grok 4.3 generally available on Bedrock, the first time an xAI model has been offered through the platform. The move makes xAI the third major independent lab available on Bedrock alongside Anthropic and OpenAI, turning Amazon's managed-model marketplace into a genuine multi-lab venue where enterprises can swap frontier models behind a single procurement and security layer.

The positioning is aggressive on price. At $1.25 per million input tokens and $2.50 per million output tokens, Grok 4.3 is the cheapest US-lab frontier reasoning model on Bedrock, paired with a 1-million-token context window and configurable reasoning effort. It runs on Mantle, Amazon's new inference engine built for price-performance, with support for tool calling, structured output and streaming.

โ€œIt runs on Mantle, Amazon's new inference engine built for price-performance, with support for tool calling, structured output and streaming.โ€

For xAI, distribution through AWS is the real prize: it instantly reaches enterprises that buy through Amazon and won't touch a standalone API. For Anthropic and OpenAI, a low-priced, capable rival on the same shelf sharpens the cost competition just as the Trump administration's restrictions complicate Anthropic's footing -- a reminder that on Bedrock, model choice increasingly comes down to price, latency and trust.

ShareXLinkedInEmail
More onAnthropic โ†’OpenAI โ†’Amazon โ†’xAI โ†’

Originally reported by xAI. Analysis and editorial commentary by Value Add Pulse.

โ† Back to Pulse

THE WIRE in your inboxโ€” Tech, startup & VC news with Trace's take. Free, no spam.

Read Next

AI

Nvidia Unveils New AI Model, Expands in Japan

Nvidia unveiled a new AI model and expanded its physical-AI partnerships in Japan, extending its push beyond chips into the software and robotics layers of the AI stack.

AI

The AI Boom Is Testing the Limits of Growth

Axios reports the AI investment boom is running into real physical and financial constraints -- power availability, chip supply and capital costs -- that could cap how fast the buildout can continue.

AI

Ex-OpenAI CTO's Thinking Machines Open-Sources Inkling

Thinking Machines, founded by former OpenAI CTO Mira Murati, open-sourced its first multimodal model, Inkling, built for low cost and resistance to censorship, and it topped Hacker News on release.

@Trace_Cohenยทt@nyvp.com