AI Trend Notifier
EN
← wiki

$ cat wiki/models/gpt-image-2-5-flare.md

GPT-Image-2.5 Flare

modelupdated 2026-09-09created 2026-09-09

Compared with

Spec

AttributeValue
DeveloperOpenAI
Released2026-09-08
Announced2026-09-08
Context windowunknown
Pricingimage input $8.00/M tokens ($2.00/M cached) · image output $30.00/M tokens · text input $5.00/M tokens ($1.25/M cached)
Licenseproprietary
AvailabilityOpenAI API
The Pricing row is the published API rate carried by two independent passes,
and it is reported as unchanged from GPT Image 2 — the quality and speed
improvement is priced at zero
(source).

Context window is unknown and this is the row that matters most here. No resolution, image-size or token-accounting figure was published, so the per-token price cannot be turned into a price per image on this wiki. A model priced per token whose token count per image is unknown is, for comparison purposes, unpriced.

Release Date

2026-09-08, alongside ChatGPT Images 2.5 and GPT-Image-2.5 Sunburst. Captured 2026-09-09 through prefetch candidate #42, +1 day (source).

Benchmarks

None published. The only comparative claim is against a sibling: higher quality than GPT Image 2 at 50% lower latency (2 passes). No score, no evaluation, no third-party measurement appears in anything read (source).

Use Cases

OpenAI positions this as the default choice for most applications. Named uses: creator content, social media, product experiences, visual search, rapid prototyping and high-volume generation (source).

The last of those is what the pricing row is for — at identical per-token rates to GPT-Image-2.5 Sunburst, the choice between the two is latency against edit control, not cost.

Compared To

  • GPT-Image-2.5 Sunburstsame published rates, different positioning. This is the unusual case on this wiki of two models from one vendor released the same day at the same price, differentiated only by generation time and precision. Nothing read quantifies either difference
  • ChatGPT Images 2.5 — the ChatGPT-side model, described as carrying "the same improvements". No statement of identity was read
  • GPT Image 2 — the baseline for both the quality and latency claims; no page on this wiki

Conflicting Reports

  • Pricing. One outlet's headline states the release comes with "2× API pricing" (datastudios.org, 2026-09-08). Its body was not read — the pass that surfaced it carried the headline only. Two independent passes carry the opposite: per-token rates identical to GPT Image 2. The Spec row above follows the two passes; the single headline is recorded here and not adopted (source)

Referenced by

Sources