Apple Foundation Models 3 vs Atria Dawn Preview

Apple · US  |  Shanghai AI Laboratory · China · Updated June 2026

Quick verdict

Pick Apple Foundation Models 3 for runs entirely on-device on iphone - no api call, no per-token cost, works offline or the 20b sparse 'core advanced' variant lifted on-device output quality from 23% to 46% user preference in apple's own testing versus the prior generation. Pick Atria Dawn Preview for best-in-class among tested models on several of shanghai ai lab's own benchmarks, including browsecomp (92.5, ahead of gpt-5.6 sol's 92.2) and deepsearchqa (96.0) or 744b-parameter moe built on a glm-5.2 base, released under a fully open mit license. Choose Atria Dawn Preview if you need self-hosting or data privacy; Apple Foundation Models 3 if you want a managed API.

Apple Foundation Models 3 (Apple, US) and Atria Dawn Preview (Shanghai AI Laboratory, China) line up two different AI ecosystems against each other — a comparison that is as much about cost philosophy and openness as raw capability. Apple Foundation Models 3 is apple's on-device model family - genuinely small context and no API access, but private, offline, and free to run on every iPhone. Atria Dawn Preview is a free, MIT-licensed 744B open-weight model from Shanghai AI Lab that beat GPT-5.6 Sol on the BrowseComp benchmark — released quietly, weights before paper. They diverge most on context window and open vs. closed weights — each quantified below from the models' real specs.

Key differences at a glance

Side-by-side specs

SpecApple Foundation Models 3Atria Dawn Preview
ProviderApple (US) Shanghai AI Laboratory (China)
ReleasedJune 2026 September 11, 2026
Context window4K (~6 pages) 256K tokens (~393 pages)
Price (in/out)Not published Open weight (self-host / free)
Open weight?No — API only Yes — self-hostable
Modalitiestext, image text
SWE-Bench VerifiedNot published Not published
MRCR v2 @ 1MNot published Not published

Who wins what

Runs entirely on-device on iPhone - no API call, no per-token cost, works offline

Apple Foundation Models 3

Atria Dawn Preview is comparatively weak here — context window is reported inconsistently across trackers — most list 256K tokens, at least one lists 1M; unconfirmed which is accurate

The 20B sparse 'Core Advanced' variant lifted on-device output quality from 23% to 46% user preference in Apple's own testing versus the prior generation

Apple Foundation Models 3

Apple Foundation Models 3 lists the 20B sparse 'Core Advanced' variant lifted on-device output quality from 23% to 46% user preference in Apple's own testing versus the prior generation among its strengths; Atria Dawn Preview does not.

Deep integration with iOS apps via the Foundation Models framework, now open to third-party LLM providers too

Apple Foundation Models 3

Apple Foundation Models 3 lists deep integration with iOS apps via the Foundation Models framework, now open to third-party LLM providers too among its strengths; Atria Dawn Preview does not.

Best-in-class among tested models on several of Shanghai AI Lab's own benchmarks, including BrowseComp (92.5, ahead of GPT-5.6 Sol's 92.2) and DeepSearchQA (96.0)

Atria Dawn Preview

Apple Foundation Models 3 is comparatively weak here — a genuinely small context window (about 4,096 tokens, shared between input and output) - far below cloud frontier models

744B-parameter MoE built on a GLM-5.2 base, released under a fully open MIT license

Atria Dawn Preview

Open weights make this possible at all — Apple Foundation Models 3 is API-only, so it cannot leave the vendor's servers.

Free to self-host — no API pricing, run entirely on your own hardware

Atria Dawn Preview

Apple Foundation Models 3 is comparatively weak here — no public API pricing since it isn't sold per-token - only usable within Apple's own platforms

Largest single-prompt input

Atria Dawn Preview

Its 256K tokens window is about 64× larger than Apple Foundation Models 3's 4K, fitting roughly 393 pages in one prompt.

Which should you pick?

Someone analysing very long documents or codebases

Atria Dawn Preview

Larger 256K tokens window fits more in one prompt.

A team with data-privacy or self-hosting needs

Atria Dawn Preview

Open weights let you run it on your own hardware; Apple Foundation Models 3 is API-only.

Anyone whose priority is runs entirely on-device on iphone - no api call, no per-token cost, works offline

Apple Foundation Models 3

It is specifically built for that.

Anyone whose priority is best-in-class among tested models on several of shanghai ai lab's own benchmarks, including browsecomp (92.5, ahead of gpt-5.6 sol's 92.2) and deepsearchqa (96.0)

Atria Dawn Preview

That is its strongest area.

An enterprise with regional data-residency rules

Apple Foundation Models 3 or Atria Dawn Preview

Origin (US vs China) affects where data is processed and which compliance regime applies — check the provider's terms for your region.

Apple Foundation Models 3: where it fits

Apple's on-device model family - genuinely small context and no API access, but private, offline, and free to run on every iPhone. Released June 2026 by Apple, it is built for runs entirely on-device on iPhone - no API call, no per-token cost, works offline, the 20B sparse 'Core Advanced' variant lifted on-device output quality from 23% to 46% user preference in Apple's own testing versus the prior generation, deep integration with iOS apps via the Foundation Models framework, now open to third-party LLM providers too, and zero marginal cost and strong privacy - nothing leaves the device.

Its trade-offs are real: a genuinely small context window (about 4,096 tokens, shared between input and output) - far below cloud frontier models, not comparable in raw capability to frontier cloud models - built for efficiency and privacy, not benchmark leadership, and no public API pricing since it isn't sold per-token - only usable within Apple's own platforms.

Atria Dawn Preview: where it fits

A free, MIT-licensed 744B open-weight model from Shanghai AI Lab that beat GPT-5.6 Sol on the BrowseComp benchmark — released quietly, weights before paper. Released September 11, 2026 by Shanghai AI Laboratory, it is built for best-in-class among tested models on several of Shanghai AI Lab's own benchmarks, including BrowseComp (92.5, ahead of GPT-5.6 Sol's 92.2) and DeepSearchQA (96.0), 744B-parameter MoE built on a GLM-5.2 base, released under a fully open MIT license, and free to self-host — no API pricing, run entirely on your own hardware.

Its trade-offs: context window is reported inconsistently across trackers — most list 256K tokens, at least one lists 1M; unconfirmed which is accurate, released quietly, with weights and an API posted to GitHub before Shanghai AI Lab's own technical report went up on arXiv days later, and a lab research preview rather than a commercial product — support and update cadence are unclear. As an open-weight model, its running cost is your own hardware rather than a per-token fee.

The bottom line for this matchup

The defining split here is open vs. closed. Atria Dawn Preview gives you weights you control — self-host it, fine-tune it, keep data in-house, pay only for hardware. Apple Foundation Models 3 gives you a managed, always-updated API with no infrastructure to run. Teams with GPUs, privacy requirements, or huge volume often favour the open model; teams that want zero ops and the latest capabilities favour the closed one. Capability is close enough that this operational question, not the benchmark, usually decides it.

Want both Apple Foundation Models 3 and Atria Dawn Preview without two subscriptions? LumiChats gives you these plus 40+ models under one ₹69/day pass (about $1/day) — draft with one, cross-check with the other.

See pricing

Frequently asked questions

Is Apple Foundation Models 3 or Atria Dawn Preview better for coding?

Public SWE-Bench figures are not available for either model, so the honest test is your own repository — run an identical real bug through both. By design, Apple Foundation Models 3 leans toward runs entirely on-device on iphone - no api call, no per-token cost, works offline while Atria Dawn Preview leans toward best-in-class among tested models on several of shanghai ai lab's own benchmarks, including browsecomp (92.5, ahead of gpt-5.6 sol's 92.2) and deepsearchqa (96.0), and that positioning usually predicts which feels better on your codebase.

Which is cheaper, Apple Foundation Models 3 or Atria Dawn Preview?

Atria Dawn Preview is open-weight, so self-hosting means no per-token fee (you pay for hardware instead), while Apple Foundation Models 3 is API-metered at Not published. For most teams without GPUs, the API model is cheaper to start; at very high volume, self-hosting can win.

Which has the bigger context window?

Atria Dawn Preview — 256K tokens vs 4K, about 64× larger. Useful only if the model actually reasons over the full window, which not all do.

Can I use both Apple Foundation Models 3 and Atria Dawn Preview together?

Yes — a multi-model platform like LumiChats gives you Apple Foundation Models 3, Atria Dawn Preview and 40+ others under one ₹69/day pass (about $1/day), so you can draft with one and cross-check with the other instead of buying two subscriptions.

Which is newer, Apple Foundation Models 3 or Atria Dawn Preview?

Atria Dawn Preview — released September 11, 2026, about 3 months after Apple Foundation Models 3.

Related comparisons

Specifications and benchmarks reflect publicly reported figures as of June 2026 and may change as providers release updates. Always verify on your own workload.