【文章标题】：The Pelican comparison grid for Astra is pretty interesting
【文章标题】：Astra的鹈鹕对比网格相当有趣
【文章正文】：
I got access to GPT-6 Astra this afternoon, so naturally I used it to generate
今天下午我获得了GPT-6 Astra的访问权限，于是很自然地用它生成了
SVGs of pelicans riding bicycles
骑自行车的鹈鹕的SVG图像
- at low, medium, high, xhigh and max reasoning levels (Astra doesn’t support reasoning=none). Then I rendered those pelicans in
- 在低、中、高、极高和最大推理级别下（Astra不支持无推理模式）。然后我将这些鹈鹕图像以
a comparison grid
对比网格的形式
with GPT-5.6 Sol, Terra, and Luna, and beyond being fun the result was surprisingly useful.
与GPT-5.6 Sol、Terra和Luna进行对比，结果不仅有趣，还出人意料地实用。
See
查看
the grid
对比网格
for full quality images. Here’s
获取完整质量的图像。这是
the transcript
生成GPT-6 Nova鹈鹕的
that created the GPT-6 Nova pelicans.
完整记录。
There are a few interesting things that stand out from this grid.
从这个对比网格中可以发现几个突出的有趣现象。
The Astra pelicans are
Astra生成的鹈鹕
much better
优秀得多
. The very best GPT-5.6-Sol pelican (I liked xhigh better than max) is still pretty clearly a bunch of abstract shapes. Every single one of the Astra pelicans, from low to xhigh, looks better than that. The Astra max one is really good.
。即便是GPT-5.6-Sol的最佳鹈鹕（我个人认为极高模式比最大模式更好），仍然明显是一堆抽象形状。而Astra从低到极高模式生成的每只鹈鹕都比它强，最大模式的鹈鹕更是堪称完美。
Astra below max still doesn’t reliably get the pelican legs on both sides of the frame.
非最大模式的Astra仍无法稳定生成画面两侧都有鹈鹕腿的图像。
In terms of cost, Astra may be around twice the price of Sol ($10/million input, $50/million output, compared to $5/$30 for Sol), but it uses significantly less tokens at each of the levels, making the prices at the different levels closer than they might otherwise be.
成本方面，Astra价格约为Sol的两倍（输入每百万次10美元，输出每百万次50美元，而Sol是5/30美元），但它在每个级别消耗的token数显著更少，使得不同级别间的实际价格差距比表面数字更小。
Astra low produces a better pelican than ANY of the GPT-5.6 Sol models at any level, for 9.55 cents. Spending 10 cents on any other model gets a much worse result.
Astra低模式仅需9.55美分就能生成比任何级别GPT-5.6 Sol都更好的鹈鹕，而其他模型花10美分得到的结果要差得多。
Look at the input token counts: Astra and Luna both used 16 input tokens, Sol and Terra used 26. That’s interesting.
注意输入token计数：Astra和Luna都只用了16个输入token，而Sol和Terra用了26个，这很有趣。
I wonder if Astra and Luna are more related to each other than OpenAI let on?
我在想Astra和Luna之间的关系是否比OpenAI公开的更为密切？
Tags:
标签：
ai
人工智能
,
、
openai
OpenAI
,
、
generative-ai
生成式AI
,
、
llms
大语言模型
,
、
pelican-riding-a-bicycle
骑自行车的鹈鹕
,
、
gpt-6-astra
GPT-6 Astra
🔗 知识库双向关联
- 🎙️ How I AI Grok Bot Grok 4.6—what’s great and what’s still_全翻译
- [[Raw_翻译_[AINews] 10 worse, 100x cheaper, 10000x faster Why Simulatio|10 worse, 100x cheaper, 10000x faster Why Simulatio_全翻译]]
- Ultra-High Interactivity on NVIDIA GPUs - TileRT InferenceX_全翻译