Skip to content
Back
Pricing briefing
PPerplexity
Perplexity
PACKAGING CHANGE

Perplexity swapped GPT-6.1 Sol for GPT-6.2 Sol in its model lineup

By , Co-founder, PricingSaaS

Published

Perplexity upgraded its included AI model from GPT-6.1 Sol to GPT-6.2 Sol.

Perplexity’s pricing page · Change detected

What changed

Perplexity's pricing page lists the AI models included across its plans in a 'Multiple AI models deliver richer answers' showcase. Pulse detected the change on Nov 18, 2026: the included model moved from GPT-6.1 Sol to GPT-6.2 Sol in that lineup card.

The 'New' label also moved: it no longer sits next to Gemini 4 Hydrogen and now marks GPT-6.2 Sol instead. The rest of the lineup — Gemini 4 Hydrogen, Claude Sonnet 5.5, Kimi K3, GLM 5.3 and Grok 4.7 — is unchanged, and Pulse recorded no change to plan prices or seat counts.

Why it matters

For a product leader, this move is less about the one model swap and more about the pace. Pulse tracked 9 changes to Perplexity's pricing page in the last 90 days. The company has made 12 real pricing changes since Pulse began tracking it on Jul 6, 2026, a pace of about one every 12 days. That cadence means the underlying model roster is a lever Perplexity pulls constantly, not a decision revisited once a quarter.

If you bundle access to third-party models as part of your own pricing, rather than charging by usage, this is the playbook to copy or avoid: Perplexity keeps its 'included models' list current without touching price or plan structure, so customers see fresh capability without a new SKU. Watch whether Perplexity keeps swapping within the same slot with a newer release from the same vendor, or starts adding net-new vendors to the lineup instead.

In context

Pulse has tracked 10 feature changes to this lineup since Jul 13, 2026 [[perplexity/2026D318]], most recently on Nov 14, 2026 when Perplexity added Gemini 4 Hydrogen in place of Gemini 3.8 Flash — the same slot that moved again four days later. Pulse tracked 354 feature changes among the 3,298 pricing changes it recorded in the last 180 days, about 10.7%, led by companies like HubSpot, Koala and Modal. Most of those are additions or promotions, like Koala's Black Friday discount, rather than a like-for-like swap. The sharper read: Perplexity isn't repricing anything — it's keeping its 'included models' pitch current by rotating in newer releases from the same vendors, and the badge now simply follows whichever model is freshest.

Common questions

What should we check before assuming this is a pricing signal?

Check whether the underlying price or usage limits moved alongside the model — Pulse shows neither did here, so read the badge and model swap as a freshness update to the plan page copy, not a repricing or a new feature launch.

Does Pulse classify this as a packaging change?

Pulse tagged it a feature change, not a price or plan event — the only on-page differences are the model name and which item carries the 'New' label, both within the AI-models showcase.

How does this compare with how other companies handle model updates?

Most companies Pulse tracks make bigger, isolated moves — a new plan, a discount, a feature launch — rather than rotating the same included-model slot repeatedly; Perplexity's swap-in-place approach is closer to how a cloud vendor patches a dependency than how SaaS companies usually touch their plans.

What people are saying

@TuankxaeX

“Perplexity 说 GPT-6.2 Astra 和 Sol 已接入它的 Search 和 Computer,并给出自己的 WANDR 基准:Astra 得 0.783,每次大约 12.54 美元,比 GPT-6 Astra 高 0.095;Sol 得 0.746,每次大约 2.65 美元,比 GPT-6.1 Sol 高 0.065。分数只差 0.037,费用差了大约 4.7 倍。这是 Perplexity 自己的基准和计价,不是 OpenAI 官方价。差距不大的时候,价钱会先被拿出来比。”

View original
@cccdickX

“Cost per task is the number most teams should watch. A 0.037 score gap at roughly 4.7x the price only pays off on the hard tasks where those last points change the outcome. Routing by task difficulty usually beats picking one model for everything.”

View original

Sources