Test record & evaluation basis

Freepik / Magnific AI Image Generator · Image generation

Free login generates a 4-image batch and exports 1024px watermark-free PNGs with fast inference and accurate sign text; requires manually disabling default 16:9 and prompt rewriting, and missed one apple.

Test date 2026-10-19Free tier + subscriptionNative output attachedMethod 2.3.1
1 · Date, environment & tier2026-10-19 · Signed-in flow; paid tier not standardized

US-West (California, AS2914)

2 · Fixture & native exportFixed English prompt (anchor_t2i_v1)

1024 × 1024 · PNG · Open native file ↗

3 · Price checks vs. test scoreOfficial pricing checked 2026-10-20

Subsequent pricing and service checks are shown separately without altering historical scores.

4 · What one test cannot proveScoped to this fixture & path

One run does not establish all complex scenes, untested paid maxima or long-term uptime.

Plans & historical pricing

Individual web subscriptions at the current Magnific destination; excludes Teams and separate API · 2026-10-20

Plan · source date
Monthly / one-time
Annual bill / period
Quota & limits
USD 20/month
USD 174/year
Monthly 20,000 general credits/month
Annual 240,000 general credits/year; valid for a year without monthly resets Rules [1]
USD 45/month
USD 405/year
Monthly 45,000 general credits/month
Annual 600,000 general credits/year; valid for a year without monthly resets Rules [1]
Pro Starter2026-10-20 ↗
USD 110/month
USD 990/year
Monthly 112,500 general credits/month
Annual 1,500,000 general credits/year; valid for a year without monthly resets Rules [1]
Pro high volume2026-10-20 ↗
USD 280/month
USD 2,520/year
Monthly 300,000 general credits/month

Monthly price before tax; credits do not establish removal counts

Annual 4,000,000 general credits/year; valid for one year, no monthly resets Rules [1]

8 plan / billing-period records, all listed; — means that period is not recorded.

[1] Annual price before VAT/local taxes, which depend on billing region. Renews annually; renewal can be canceled in subscription settings.

Both individual Pro volumes and all annual charges are verified. Auto-refill was not enabled; the public pricing page does not quote top-up unit prices, free removal limits or per-removal credit cost, so those remain unconfirmed. Prices exclude VAT/local taxes.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancel auto-renewal anytime in account subscription settings

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

Official subscriptions and annual checkout ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗ · Capture 5 ↗ · Capture 6 ↗

Pricing recorded on the test date · 2026-10-19

Free daily logged-in allowance; paid tiers include Essential / Premium / Premium+ (no purchase in this run; historical prices not reverified)

Freemium-Login-Required (daily free logged-in allowance + tiered subscriptions)

Historical context only; not current purchase guidance.

Delivery & access

Free delivery record
1024×1024 (Unwatermarked native PNG, 1.51 MB)
Attached file dimensions
1024 × 1024 · PNG

Read directly from the attached file; this does not establish free download availability at that size.

Watermark
Zero watermark (clean 1024×1024 PNG export)
Free allowance
Requires free login to use daily free allowance; default Auto model generates 4 candidates per batch and exports unwatermarked PNGs
Account requirements
Free Google / Apple / Email login (automatically resumes pending generation after sign-in)
Maximum specification record
1024×1024 (Unwatermarked Native PNG · Auto Model 4-Image Batch)

May include advertised capabilities; this does not establish a tested paid tier.

Batch & API
Generates 4 candidates per batch by default, with in-workspace handoff to Upscale, Relight, and Mystic models
Manual refinement
Not recorded or not verified

Tested result & sample

Under canonical square settings (explicit 1:1, AI prompt OFF), Freepik / Magnific’s Auto model renders a 4-image 1024×1024 batch in 5.76s: the "REALTEST 2026" sign text is 100% accurate (25/25) with rich brass texture, though it generates 2 red apples instead of 3 and obscures the thumb behind the sign (56.0/65).

Fixture
anchor_t2i_v1
Recorded environment
US-West (California, AS2914)
Network-observed server wait
5.76 s
End-to-end reference
6.12 s

Network wait includes round trips, queuing and polling error, not pure inference time. Local timings depend on the device; repetition counts are not standardized in legacy records.

Experience, rules & trust

Disclosure
Prompts for login after clicking generate; workspace defaults to 16:9 and AI prompt rewriting unless manually switched to 1:1 and OFF
Interaction
Free download delivers a clean 1024×1024 native PNG without watermarks or credit-card walls; however, the workspace defaults to 16:9 aspect ratio with "AI prompt" rewriting enabled, requiring manual adjustment to 1:1 and turning "AI prompt" OFF for faithful square prompt adherence.
Entity record
Freepik Company S.L. / Magnific AI
Origin
Málaga, Spain
Domain history record
2010–Present (acquired Magnific AI in 2024)
Third-party tracking record
Not recorded or not verified
External information
Not recorded or not verified
Processing
Freepik entry redirecting to Magnific parallel cloud rendering pipeline (start-tti-v2 + 4 parallel ak-data.magnific.com/app/api/render/v4 requests; 5.76s pure server time)
Retention statement
Generations are stored in the logged-in cloud gallery; free-tier deletion/privacy terms were not fully reverified in this run

A recorded statement does not verify deletion or all data-processing behavior.

The former Freepik feature and pricing pages currently redirect to Magnific; historical test identity and date are preserved.

Verifiable audit trail & evidence

Four SHA-256-verified stages link the benchmark fixture, same-run interface screenshots, native exported artifact and archived pricing sources.

Stage A · Benchmark fixture Fixed English prompt

anchor_t2i_v1 · Inspect fixture & file inspector →

Stage B · Same-run screenshots 2 flow screenshots archived

Captures entry rules, result view and export disclosure.

Stage C · Native exported file 1024 × 1024 · PNG

Open native exported file ↗

Stage D · Archived pricing sources 2 official source snapshots

Official plans and usage terms (snapshot ↗) · Official subscriptions and annual checkout (snapshot ↗)

Screenshots, native output and official claims support different conclusions. Evidence grades are not product safety certifications.

Evaluation basis & full scores · Overall 85.9/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 81/100 · 30%

CriterionScoreRationaleBasis
Pricing transparency26 / 35 Clicking generate on the Freepik AI Image Generator landing page redirects to the Magnific workspace (magnific.com/app/ai-image-generator) and prompts unauthenticated visitors to sign in; after free Google sign-in, the pending generation resumes automatically using the daily free allowance to produce a 4-image batch and export an unwatermarked 1024×1024 PNG (though login requirement and exact daily credit rules are not disclosed upfront on the landing hero). Evidence B →
Interaction restraint23 / 30 Free download delivers a clean 1024×1024 native PNG without watermarks or credit-card walls; however, the workspace defaults to 16:9 aspect ratio with "AI prompt" rewriting enabled, requiring manual adjustment to 1:1 and turning "AI prompt" OFF for faithful square prompt adherence. Evidence A →
Background & reputation19 / 20 Operated by Málaga-based Freepik Company S.L. (founded 2010; acquired Magnific AI in 2024), one of the world’s largest design asset and visual AI platforms. Evidence B →
Engineering & privacy13 / 15 Seamless cross-domain Google OAuth handoff that preserves and resumes the pre-login prompt, rendering 4 parallel candidates via ak-data.magnific.com/app/api/render/v4. Evidence A →

Capability · 88/100 · 70%

CriterionScoreRationaleBasis
Task quality56 / 65 Text Rendering (25.0/25: Exact two-line white "REALTEST 2026" text on the wooden sign with 100% spelling accuracy (25/25)) · Count & Color Binding (15.0/20: 2 blue ceramic mugs on the right are accurate without color bleeding, but only 2 red apples are generated on the left instead of the requested 3 (15/20)) · 5-Finger Hand & Structure (16.0/20: Vintage brass robotic hand holding the sign shows 4 long articulated fingers in front with the thumb obscured behind the board; solid 1:1 composition and warm lighting (16/20)) Evidence A →
Deliverable specifications20 / 20 Clean unwatermarked export (8/8); exports the first candidate from the standard 1:1 batch as a native 1024×1024 RGB PNG (1,585,669 bytes, watermark-free) (12/12) Evidence A →
Processing speed12 / 15 2026-10-21 single-sample network timing under standard conditions (Auto model, explicit 1:1, AI prompt OFF): pure server inference from POST /app/api/start-tti-v2 across 4 parallel POST /app/api/render/v4 calls to POST /app/api/v2/ai/simulate-generation completes in 5.76s (end-to-end 6.12s) (12.0/15) Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Revisions & corrections

Prepare a correction or update →
Benchmark evidence preview

Open original ↗