$ diff gpt-image-2-5-flare hy4-preview
GPT-Image-2.5 Flare vs Hy4 preview
Values come from GPT-Image-2.5 Flare and Hy4 preview, where each is cited to its source. This page states no benchmark result and ranks neither model — it puts two published specifications next to each other. Where a lab has not published a figure, the row says so rather than guessing.
OpenAIGPT-Image-2.5 Flare
- context
- —
- weights
- closed
- $/M in
- $30
- $/M out
- $2
- context
- 1M
- weights
- open
- $/M in
- $0.834
- $/M out
- $2.501
What actually differs
- Input price
- Hy4 preview at $0.834/M against $30/M — 36.0× cheaper to feed.
- Output price
- GPT-Image-2.5 Flare at $2/M against $2.501/M — 1.3× cheaper to generate. Output dominates the bill on most agentic workloads, where the model writes far more than it reads.
- Weights
- Hy4 preview publishes weights (Apache-2.0); GPT-Image-2.5 Flare is API-only. That decides self-hosting, air-gapped deployment and fine-tuning before any capability question does.
- Recency
- GPT-Image-2.5 Flare shipped 11 days after Hy4 preview (2026-09-08 vs 2026-08-28).
Full spec
| Attribute | GPT-Image-2.5 Flare | Hy4 preview |
|---|---|---|
| Developer | OpenAI | Tencent |
| Released | 2026-09-08 | 2026-08-28 |
| Context window | not recorded | over 1,000,000 |
| Pricing | image 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) | $0.834/M input · $2.501/M output |
| License | proprietary | Apache-2.0 |
| Availability | OpenAI API | Hugging Face (BF16 + FP8), ModelScope, GitCode, CNB, Tencent Cloud TokenHub, OpenRouter, WorkBuddy, CodeBuddy, Yuanbao, ima |