Maggie image tests

Named-model realism comparison

FLUX.2 Max · 1080p requested

fal-ai/flux-2-max/edit

$0.16/image

$0.16/imagerecorded API usage · Reported API usage per delivered image; excludes taxes/funding fees. Invoice not reconciled.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Same original fictional persona reference and shared head-on home-camera brief; compare photographic realism and identity separately.

Why another revision: Earlier likeness-preserving attempts still appeared synthetic; test three named models.

Reference: Same original fictional persona reference; original reference remains private. Editorial summary; human votes are separate.

No published AI review.

Named-model realism comparison

OpenAI Sunburst · max

openai/gpt-image-2.5/sunburst/edit

$0.1718/image

$0.1718/imagerecorded API usage · Reported API usage per delivered image; excludes taxes/funding fees. Invoice not reconciled.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Same original fictional persona reference and shared head-on home-camera brief; compare photographic realism and identity separately.

Why another revision: Earlier likeness-preserving attempts still appeared synthetic; test three named models.

Reference: Same original fictional persona reference; original reference remains private. Editorial summary; human votes are separate.

No published AI review.

Named-model realism comparison

Google Nano Banana Pro · 2K

fal-ai/nano-banana-pro/edit

$0.15/image

$0.15/imagerecorded API usage · Reported API usage per delivered image; excludes taxes/funding fees. Invoice not reconciled.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Same original fictional persona reference and shared head-on home-camera brief; compare photographic realism and identity separately.

Why another revision: Earlier likeness-preserving attempts still appeared synthetic; test three named models.

Reference: Same original fictional persona reference; original reference remains private. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Original likeness · new realism attempt

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Return to the original fictional persona reference while keeping a low front microphone and a natural home-camera brief.

Why another revision: The preceding text-only attempt changed the character’s appearance too much. This attempt restores the original likeness; realism and identity acceptance remain pending.

Reference: Original fictional persona board; appearance reference only. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Fresh scene without image reference

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Use a text-only generation to avoid carrying over reference texture and styling.

Why another revision: Removing the image reference changed the facial likeness too much. Realism and identity are still unresolved.

Reference: No image reference. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Fresh scene from persona reference

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Generate a new scene using the original fictional persona board for appearance.

Why another revision: The result felt too polished and staged; try an ordinary home-recording treatment.

Reference: Original fictional persona board; appearance reference only. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Rebuild from earlier reference

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Rebuild the low-mic scene from the earlier head-on reference instead of the edited chain.

Why another revision: A texture concern was still visible in the saved image; try a fresh generation.

Reference: Earlier head-on reference. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Targeted skin correction

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Try to remove repeated skin texture while retaining natural age lines.

Why another revision: The texture concern remained unresolved, so return to an earlier reference.

Reference: Soft-background target and prior relaxed-pose reference. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Softened background

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Add optical background softness to separate the presenter from the room.

Why another revision: Patterned, etched-looking facial texture remained distracting; investigate skin detail.

Reference: Relaxed-pose image edit. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Relaxed pose, lower front mic

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Move the mic below the neckline and sit farther back in the same setup.

Why another revision: Test mild background softness while keeping the face and microphone clear.

Reference: Cropped-stand image edit. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Stand cropped below frame

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Remove the extra tabletop and crop the microphone support below the frame.

Why another revision: Lower the mic in front of the torso and relax the forward-leaning pose.

Reference: Boom-mic image edit. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Desk stand with visible base

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Place the Shure-style mic on a weighted tabletop stand.

Why another revision: The added angled foreground tabletop made the furniture geometry less convincing.

Reference: Boom-mic image edit. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Compact podcast desk mic

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Compare a compact metal-grille microphone on a tabletop stand.

Why another revision: Continue with the Shure-style microphone while adjusting its support.

Reference: Earlier head-on reference. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

Shure-style boom mic

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Compare a foam-covered podcast mic supported by a boom arm.

Why another revision: Try a desk-supported microphone for a simpler home setup.

Reference: Earlier head-on reference. Editorial summary; human votes are separate.

No published AI review.

Microphone & realism iterations

First visible desk mic

ChatGPT built-in image tool · exact backend unknown

Price unknown

Price unknownrecorded API usage · Built-in tool did not expose a USD charge; not assumed free.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Your vote

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Details & revision notes

Add a visible microphone to test whether the polished narration fits the recording setup.

Why another revision: Explore a more familiar podcast microphone and a less dominant placement.

Reference: Earlier head-on reference. Editorial summary; human votes are separate.

No published AI review.

Compare original source image ↗

HEAD-ON EDITS

Straight-on framing

Head-on edit of revision 3

ChatGPT image edit · exact backend not exposed

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Head-on edit of revision 4

ChatGPT image edit · exact backend not exposed

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Compare original source image ↗

About the head-on edits

Miguel liked revisions 3 and 4 and requested frontal framing. These edits use those exact Nano Banana Pro sources. ChatGPT’s built-in edit tool did not expose its backend or cost. AI scores and Miguel’s acceptance stay separate.

ORIGINAL + EIGHT REVISIONS + THREE CANDIDATES

Earlier images

Original phone image

OpenAI image tool · exact backend unknown

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 3/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 3/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Image revision 1

Nano Banana Pro

Front-facing; phone on the desk. Framing note, not a score.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Image revision 2

Nano Banana Pro

Front-facing; similar setup to revision 1. Framing note, not a score.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Image revision 3

Nano Banana Pro

Low, wide side angle; face turned toward the window. Framing note, not a score.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Image revision 4

Nano Banana Pro

Face-forward, but an off-center close camera and ceiling in view. Framing note, not a score.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Image revision 5

Nano Banana Pro

Three-quarter desk angle; monitor visible. Framing note, not a score.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 3/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Image revision 6

Nano Banana Pro

Three-quarter desk angle; monitor removed. Framing note, not a score.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 2/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 2/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 3/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Image revision 7

Nano Banana Pro

Slightly wider desk shot; face turned toward camera. Framing note, not a score.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 2/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Image revision 8

Nano Banana Pro

Head-on desk shot. Framing note, not a score.

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 3/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Candidate A

OpenAI image tool

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 3/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Candidate B

Nano Banana Pro

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Candidate C

OpenAI image tool

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 2/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

THE $10 ROUND

Two photographic rebuilds.

Photographic reconstruction

GPT Image 2.5 Sunburst · max quality

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Less cosmetic finish

GPT Image 2.5 Sunburst · max quality

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
still_image · 2026-10-16. Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

THE $20 COMPARISON

Five image models.

The agent chose two sources for six speaking clips; Miguel had not approved them.

Google Nano Banana2.1

5504×3072 · high thinking ·4K requested

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 3/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 3/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Google Nano Banana Pro

5504×3072 · 4K requested

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · Unknown score
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 2/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

OpenAI GPT Image2.5 Sunburst

3840×2160 · max quality

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 3/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · Unknown score
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

FLUX.2 Max

2560×1440 · 2560×1440 requested

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 4/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 2/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 4/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.

Qwen Image Edit2511

2560×1440 · 50steps · no acceleration · open-weight comparator

AI judgesFive-judge review pending0/5

A panel score needs five compatible numeric scores. Unknown results are not zero. AI diagnostics; community votes and Miguel’s verdict stay separate.

ClaudePending—
anthropic/claude-sonnet-5.5
GeminiPending—
google/gemini-4-argon
GrokPending—
x-ai/grok-4.7
QwenPending—
qwen/qwen3-vl-235b-a22b-instruct
MistralPending—
mistralai/mistral-large-4-0

Video frame reviews cover sampled visuals; they do not certify audio, lip sync or every frame.

Earlier reviews · 3

google/gemini-3.8-flash · 4/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

openai/gpt-5.6-sol · 2/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

qwen/qwen3-vl-235b-a22b-instruct · 3/4
One resized still preview; motion/audio not reviewed · . Earlier model/rubric; excluded from this five-judge score.

Voting help

If your browser blocks embedded sign-in, open this review separately ↗. Votes and Credits use the same Navo account.