o10

New lab models, August 2026

Three major-lab releases in the week of 10–13 August 2026: xAI/Cursor Grok 4.6, Google Gemini 3.7 Flash, and Meta Muse Glimmer. Ranked here by list input price so you can see the spend spread before evals.

August 2026 lab wave: Muse Glimmer (Meta, 30B Apache 2.0), Gemini 3.7 Flash (Google workhorse with intro pricing), and Grok 4.6 (xAI and Cursor agent flagship).

Three major-lab releases in the week of 10–13 August 2026: xAI/Cursor Grok 4.6, Google Gemini 3.7 Flash, and Meta Muse Glimmer. Ranked here by list input price so you can see the spend spread before evals.

  1. Muse Glimmer . $0.35/M input · C1 pricing
  2. Gemini 3.7 Flash . $0.75/M input · C2 pricing
  3. Grok 4.6 . $2/M input · C3 pricing
Start hereQuick overview

How to use this index

Which major labs shipped models in August 2026?

Meta Superintelligence Labs released Muse Glimmer (open 30B, around 10 August). xAI and Cursor released Grok 4.6 on 12 August. Google released Gemini 3.7 Flash on 13 August. All three are in the o10 catalog with hosts, list prices, and eval-gated routing bands.

Overlay snapshot 2026-08-14.

How do August 2026 launch prices compare?

Hosted Muse Glimmer is about $0.35/M input. Gemini 3.7 Flash intro is $0.75/M input through 31 December 2026. Grok 4.6 is $2/M input. Muse Glimmer is $0 on a local GPU. Those are list prices, not eval winners.

Lab and host list prices, August 2026.

FAQFrequently asked questions

Common questions

What is Grok 4.6?

xAI and Cursor's 12 August 2026 agent model. $2/$6 per 1M tokens. Details: /models/grok-4-6.

What is Gemini 3.7 Flash?

Google's 13 August 2026 coding/agent Flash model. Intro $0.75/$3.75 per 1M through 31 Dec 2026. Details: /models/gemini-3-7-flash.

What is Muse Glimmer?

Meta's 30B Apache 2.0 local-agent model. Hosted ~$0.35/$1.50 per 1M; local $0. Details: /models/muse-glimmer.

o10Set the envelope. o10 holds it.

See what you're overpaying.

Paste a week of traffic. Get the number that books the audit.

See what you're overpaying