top of page
Search

Perplexity AI: is the latest model included in standard subscriptions?

ree

Perplexity AI continues to evolve quickly, releasing frequent model upgrades and expanding its Pro, Enterprise, and API offerings. However, the availability of the latest models depends on your subscription tier. While Pro and Enterprise users receive early access to new checkpoints, Free-tier users experience delayed updates. Here we break down model availability, subscription differences, and rollout timelines as of August-September 2025.



The free plan delivers solid access but with delayed upgrades.

Perplexity’s Free tier offers access to a capable AI assistant but does not include instant updates when new model checkpoints launch.

  • Default models:

    • pplx-7b-online-2025-07 (lightweight, optimized for speed).

    • pplx-70b-online-2025-04 (large reasoning model).

  • Update frequency: Free users receive checkpoint upgrades quarterly. The next scheduled refresh is in October 2025, adding the pplx-70b-online-2025-07 release.

  • Context window: Up to 48,000 tokens per query.

  • Usage limits:

    • 50 queries per day.

    • Maximum 5 requests per minute.

For casual users, the free plan covers most daily needs but lacks priority access to cutting-edge model improvements.



Pro subscribers get faster access to the newest models.

Perplexity’s Pro plan—priced at $20/month—is designed for power users, developers, and researchers who need up-to-date models and higher performance limits.

  • Default models:

    • pplx-70b-online-2025-07 (latest 70B reasoning model).

    • pplx-2.5-mixtral-48b (optional beta model via Labs toggle).

  • Early rollout: Pro subscribers typically receive new checkpoints within 48 hours of release.

    • The 2025-07 checkpoint was available to Pro users on 29 July, two days after Enterprise deployment.

  • Context window: 100,000 tokens per request (expanded from 70k on 1 August 2025).

  • Usage limits:

    • 800 queries per day.

    • Burst rate of 15 requests per minute.

  • Labs features: Pro users can experiment with upcoming capabilities like the Mixtral-48B beta, which introduces multi-query reasoning paths.

For individuals needing faster model access, Pro offers a significant advantage over the free tier.



Enterprise plans unlock instant updates and early previews.

Perplexity’s Enterprise tier provides premium organizations with real-time access to the latest model checkpoints, advanced context handling, and optional experimental builds.

  • Default model: pplx-70b-online-latest (automatically updated on release).

  • FastTrack channel: Enables nightly checkpoint deployments, giving early access to test models before they reach Pro users.

  • Context windows: Configurable up to 200,000 tokens for complex workloads.

  • Pooled usage limits:

    • 2 million queries/day by default.

    • Upgradeable to 5 million/day with enhanced contracts.

  • Integration benefits:

    • Priority routing for faster response times.

    • Admin tools to pin model versions for compatibility testing.

    • Rollback options in case new checkpoints affect results.

Enterprise customers gain the highest degree of control, reliability, and priority performance—ideal for teams requiring data-intensive operations.


API access always uses the latest stable model.

For developers integrating Perplexity into their own tools, the /v1/chat API guarantees automatic access to the newest public checkpoint:

  • Pricing: $15 per million tokens.

  • Default model alias: pplx-latest always points to the most recent stable release.

  • Customization: Developers can pin older model IDs if compatibility is required.

  • Throughput limits:

    • 10 queries per second by default.

    • Upgradeable up to 50 QPS upon request.

This approach ensures that custom apps immediately benefit from Perplexity’s latest reasoning improvements without waiting for subscription-based rollouts.


Model availability and update timelines.

Tier

Monthly Price

Default Models

Context Window

Update Speed

Usage Limits

Free

$0

pplx-7b-online-2025-07 · pplx-70b-online-2025-04

48k tokens

Quarterly (Oct 2025 next)

50 queries/day

Pro

$20

pplx-70b-online-2025-07 · Mixtral-48B beta

100k tokens

48 h after release

800 queries/day

Enterprise

Custom

pplx-70b-online-latest

Up to 200k tokens

Instant + nightly FastTrack

2M/day default; upgradable to 5M

API

Pay-go

pplx-latest alias

100k tokens

Immediate

10 QPS (up to 50)


Choosing the right plan for latest-model access.

Scenario

Recommended Tier

Why

Occasional research & Q&A

Free

Covers basic queries but delays new model access.

Up-to-date access for individuals

Pro

Checkpoints roll out within 48 hours; Labs beta features included.

Enterprise-scale AI workloads

Enterprise

Instant updates, configurable context, and higher daily limits.

Building apps & integrations

API

Always synced to the latest stable release, with scalable QPS options.



Perplexity AI’s latest checkpoints are rolled out first to Enterprise users, followed by Pro subscribers within 48 hours, while Free-tier users receive updates only during scheduled quarterly refreshes. Developers using the API always have immediate access to new stable models, making it the most reliable option for cutting-edge integrations.


____________

FOLLOW US FOR MORE.


DATA STUDIOS


bottom of page