OpenAIImage modelActive
GPT-Image-2.6 Flare
Summary
GPT-Image-2.6 Flare is the speed-oriented variant from OpenAI's Images 2.6 launch on November 4, 2026. The model generates and edits images for quick iteration; at low quality settings, OpenAI cites generation times from six seconds.
It can be reached worldwide in ChatGPT and Codex on all subscription tiers and by API under the ID gpt-image-2.6-flare. Inputs can be text and images, and the output is an image. Prices per million tokens are $5 for text, $8 for image input, and $30 for image output. For detailed compositions there is GPT-Image-2.6 Sunburst.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 20 published benchmark results
Page 1 of 2
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
BudgetPixel AI Image Arena: Architecture & interiors Elo rating (Show measurement, test conditions, and source)
- Source value
- 1482
- Score
- 1482
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Architecture, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Architecture)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category architecture. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=1 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: Concept art & fantasy Elo rating (Show measurement, test conditions, and source)
- Source value
- 1196
- Score
- 1196
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Concept, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Lower confidence bound
- 1056
- Upper confidence bound
- 1500
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Concept)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category concept. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=6 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: Illustration & anime Elo rating (Show measurement, test conditions, and source)
- Source value
- 1525
- Score
- 1525
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Illustration, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Illustration)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category illustration. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=11 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: Landscape & nature Elo rating (Show measurement, test conditions, and source)
- Source value
- 1520
- Score
- 1520
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Landscape, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Landscape)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category landscape. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=12 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: overall blind-preference rating (Show measurement, test conditions, and source)
- Source value
- 1186
- Score
- 1186
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Overall, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Lower confidence bound
- 1040
- Upper confidence bound
- 1500
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Overall)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category overall. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=47 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: Photorealism Elo rating (Show measurement, test conditions, and source)
- Source value
- 1558
- Score
- 1558
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Photorealism, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Photorealism)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category photorealism. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=6 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: Portrait & fashion Elo rating (Show measurement, test conditions, and source)
- Source value
- 1514
- Score
- 1514
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Portrait, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Portrait)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category portrait. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=3 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: Product & commercial Elo rating (Show measurement, test conditions, and source)
- Source value
- 1585
- Score
- 1585
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Product, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Product)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category product. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=5 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: Typography Elo rating (Show measurement, test conditions, and source)
- Source value
- 1518
- Score
- 1518
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Typography, live leaderboard snapshot 2026-11-05; provisional collecting row
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Typography)
- Status
- superseded
- Retrieved at
- 2026-11-05
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category typography. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=3 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this model row as collecting. Pair comparisons are not independent tasks; missing confidence intervals remain unknown.
BudgetPixel AI Image Arena: Architecture & interiors Elo rating (Show measurement, test conditions, and source)
- Source value
- 1440
- Score
- 1440
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Architecture, live leaderboard snapshot 2026-11-06; provisional collecting row
- Direction
- Higher is better
- Lower confidence bound
- 1173
- Upper confidence bound
- 1678
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Architecture)
- Status
- active
- Retrieved at
- 2026-11-06
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category architecture. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=1 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this exact source row as collecting. Elo and any displayed confidence interval remain provisional until collection is complete. Votes are pairwise comparisons, not independent tasks; sampleCount remains omitted.
BudgetPixel AI Image Arena: Concept art & fantasy Elo rating (Show measurement, test conditions, and source)
- Source value
- 1505
- Score
- 1505
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Concept, live leaderboard snapshot 2026-11-06; provisional collecting row
- Direction
- Higher is better
- Lower confidence bound
- 1262
- Upper confidence bound
- 1841
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Concept)
- Status
- active
- Retrieved at
- 2026-11-06
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category concept. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=6 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this exact source row as collecting. Elo and any displayed confidence interval remain provisional until collection is complete. Votes are pairwise comparisons, not independent tasks; sampleCount remains omitted.
BudgetPixel AI Image Arena: Illustration & anime Elo rating (Show measurement, test conditions, and source)
- Source value
- 1581
- Score
- 1581
- Metric
- Bradley-Terry Elo rating from blind human pairwise image preferences; 1500 is the platform average
- Unit
- elo
- Benchmark version
- BudgetPixel AI Image Arena, Illustration, live leaderboard snapshot 2026-11-06; provisional collecting row
- Direction
- Higher is better
- Lower confidence bound
- 1389
- Upper confidence bound
- 1768
- Source type
- independent-evaluator
- Evaluator
- BudgetPixel AI Image Arena public leaderboard CSV (Illustration)
- Status
- active
- Retrieved at
- 2026-11-06
- Methodology
- Independent evaluation. Exact public CSV model ID gpt-image-2.6-flare, display_name GPT Image 2.6 Flare, maker OpenAI, category illustration. BudgetPixel uses the same prompts and default model settings, refits blind pairwise votes with Bradley-Terry, and reports an Elo scale centered on 1500. Source status is collecting; no completed rank is published. votes=11 counts pair comparisons, not independent prompts or people. Provider serving revision is not published.
- Context
- Provisional arena estimate. BudgetPixel still marks this exact source row as collecting. Elo and any displayed confidence interval remain provisional until collection is complete. Votes are pairwise comparisons, not independent tasks; sampleCount remains omitted.
Page 1 of 2
All specifications
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
- Availability
- Available globally in ChatGPT and Codex across all subscription tiers, as well as through the API.
- Source: Availability
API
Sources and data date
Show all 14 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
No active result from the reviewed external evaluators can currently be attributed to this model. Other source evidence remains available below.
Without vendor claims and historical or disputed results; those stay labelled in the single values.