{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-09-28T06:03:00.468Z","headline":"OpenRouter 实测 20 个图像生成模型的成本、编辑与质量","description":"OpenRouter 对其路由的 20 个图像生成模型用相同提示词实测计费，单张图片成本在 $0.006 到 $0.134 之间，相差 22 倍。","url":"https://www.aioga.com/news/cmu74ww1u0qfxrowkb6qtruwz/","mainEntityOfPage":"https://www.aioga.com/news/cmu74ww1u0qfxrowkb6qtruwz/","datePublished":"2026-09-18T00:00:00.000Z","dateModified":"2026-09-18T00:00:00.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://openrouter.ai/blog/insights/image-generation-models-compared","https://aihot.news/items/cmu74ww1u0qfxrowkb6qtruwz"],"canonicalUrl":"https://www.aioga.com/news/cmu74ww1u0qfxrowkb6qtruwz/","directAnswer":{"@type":"Answer","text":"OpenRouter 对其路由的 20 个图像生成模型使用相同提示词、相同尺寸进行实测，并记录每次调用的计费、返回结果与可接受参数。测试显示，单张图片成本介于 0.006 美元与 0.134 美元之间，差异约为 22 倍。","url":"https://www.aioga.com/news/cmu74ww1u0qfxrowkb6qtruwz/","dateCreated":"2026-09-18T00:00:00.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"openrouter.ai source article","url":"https://openrouter.ai/blog/insights/image-generation-models-compared","datePublished":"2026-09-18T00:00:00.000Z","provider":{"@type":"Organization","name":"openrouter.ai","url":"https://openrouter.ai/blog/insights/image-generation-models-compared"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.news/items/cmu74ww1u0qfxrowkb6qtruwz","datePublished":"2026-09-18T00:00:00.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.news/items/cmu74ww1u0qfxrowkb6qtruwz"}}],"aggregationSource":"OpenRouter：Announcements（RSS）","originalPublisher":{"name":"openrouter.ai","url":"https://openrouter.ai/blog/insights/image-generation-models-compared"},"geoDeepAnswer":null,"article":{"id":"cmu74ww1u0qfxrowkb6qtruwz","slug":"cmu74ww1u0qfxrowkb6qtruwz","url":"https://www.aioga.com/news/cmu74ww1u0qfxrowkb6qtruwz/","title":"OpenRouter 实测 20 个图像生成模型的成本、编辑与质量","title_en":"","summary":"OpenRouter 对其路由的 20 个图像生成模型用相同提示词实测计费，单张图片成本在 $0.006 到 $0.134 之间，相差 22 倍。","source":"OpenRouter：Announcements（RSS）","sourceUrl":"https://openrouter.ai/blog/insights/image-generation-models-compared","aiHotUrl":"https://aihot.news/items/cmu74ww1u0qfxrowkb6qtruwz","publishedAt":"2026-09-18T00:00:00.000Z","category":"行业动态","score":72,"selected":true,"articleBody":["We route 52 image generation models through one API：https://openrouter.ai/models?output_modalities=image. We ran the same prompt through 20 of them and recorded what each call billed, what it returned, and what it accepts.","Cost is the hard part, because these models are not sold in the same unit. FLUX.2 is priced per megapixel. Gemini, OpenAI, and Microsoft MAI are priced per token. Grok, Recraft, Riverflow, Qwen, and Seedream are priced per image, at a rate that changes with the quality and resolution you ask for. Each model page is correct in its own unit, so you cannot compare two of them by reading their prices side by side.","We put them on the same footing. The 20 models we tested all got the same prompt at the same size, and we read the usage.cost field from each response, which is the amount we charged.","Every price on this page came from a real generation, measured on 11 September 2026. Prices and models change often, so treat any figure more than a month old as a hint rather than a budget line. You can re-pull the current state with GET /api/v1/images/models ：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#via-the-image-models-api for accepted parameters and GET /api/v1/images/models/{model}/endpoints ：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#per-endpoint-records for pricing with the unit named.","We sent each of these 20 models the prompt “A studio packshot of a glass bottle, hard key light, no text.” with aspect_ratio: \"1:1\" and no other settings, then read usage.cost off the response. The billing unit column is the unit field from each endpoint’s pricing record.","Figure 1. The same prices, sorted. The three billing units are mixed all the way up the range, so the unit does not tell you where a model lands in it.","Three things to know before you budget from this table.","On OpenAI models, quality changes your bill more than any other setting. We ran gpt-image-2 with the same prompt at the same 1024x1024 size with quality unset, set to low , and set to high .","Leaving quality unset gave us the same token count and bill as low . Set quality explicitly when the image matters, and treat a per-image price for an OpenAI model as incomplete unless it names the quality tier. The high call also took 123 seconds to return against 12 seconds for the default call.","On a per-megapixel model, the frame you ask for changes the bill, but by less than the pixel count suggests. We ran black-forest-labs/flux.2-klein-4b：https://openrouter.ai/black-forest-labs/flux.2-klein-4b at two shapes. The 21:9 call returned 2400x1024, which is 2.46 megapixels against 1.05 for the square. We were billed $0.014 for the square and $0.016 for the wide crop.","Each model publishes a rate on its page and on the API’s endpoints route. Those rates are a planning number, and the response is what you pay. On 11 September 2026, sourceful/riverflow-v2.5-pro listed $0.13 per image at its default resolution and billed us $0.064. black-forest-labs/flux.2-flex listed $0.06 per megapixel and billed us $0.05 for a one-megapixel image. black-forest-labs/flux.2-max listed $0.07 and billed us exactly $0.07. krea/krea-2-medium published no pricing record on its endpoint and billed us $0.03.","When the two disagree, read usage.cost ：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#response-format rather than multiplying a listed rate by an estimate.","We bill a completed image request in full and charge nothing for one that fails. The Image API billing documentation：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#billing-and-cancellation describes which outcomes are billed.","Price is one axis. These are the others, and two of them rule models out before cost matters at all. The values come from each endpoint’s supported_parameters on 11 September 2026.","A reference image is one you send with the request for the model to work from, either an image you want edited or an example of the style or product you want matched. The limit on how many you can send is a hard ceiling. Grok takes three and Gemini takes fourteen, so a six-image brand kit rules Grok out before you compare anything else.","The output modalities column is the second axis. Most of these models take text and images in and return only image bytes. The Gemini image models and OpenAI’s gpt-5-image models can also return written text in the same reply, which matters if you want the model to explain what it made. We tested how reliably they do that further down.","Re-pull this table for any model with GET /api/v1/images/models ：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#via-the-image-models-api, which publishes each model’s accepted parameters and their ranges. A parameter that is absent from supported_parameters is not supported on that endpoint.","seed is the number you send to make a result repeatable. It is missing from every Gemini, Grok, OpenAI, Riverflow, Recraft, and MAI image model on our catalog. It exists on FLUX.2, Seedream, Qwen, and Krea.","We ran black-forest-labs/flux.2-klein-4b twice with the same prompt and seed: 424242 . The two responses were identical byte for byte, and both billed $0.014. We only tested Klein, so treat the other seed-capable models as untested rather than assuming they behave the same way. For comparison, two sourceful/riverflow-v2.5-fast calls with no seed parameter returned two different images from the same request body.","We show Design Arena：https://www.designarena.ai/leaderboard/image ratings on each model’s page. When we pulled the current ratings through our benchmarks API：https://openrouter.ai/docs/api/api-reference/benchmarks on 11 September 2026, sourceful/riverflow-v2.5-pro had the highest rating on both the image board and the image editing board. Microsoft’s MAI Image 2.6 models and Google’s Gemini image models followed on the image board, and black-forest-labs/flux.2-klein-4b had the lowest rating of the models listed on both boards.","Those ratings come from head-to-head votes between models, which Design Arena turns into an Elo rating, a score that rises when a model wins a matchup and falls when it loses.","Two caveats. The Gemini ratings sit on the -preview versions of those slugs rather than the general release slugs we link on this page. And Elo tells you what voters preferred on their prompts, not yours. The top-rated model billed $0.064 per image in our run and the lowest-rated one billed $0.014. Shortlist from the board, then decide from your own prompts.","We sent one prompt to six models across the price range and put the bill under each image. We asked for a matte black coffee bag with two lines of text printed on it.","A studio packshot of a matte black coffee bag on a plain white background, hard key light. Printed across the front of the bag in clean sans-serif capitals are the words OPENROUTER ROASTERS, and beneath that in smaller capitals the words SINGLE ORIGIN.","We picked readable words because they are easy to grade. Either the bag says what you asked for or it does not.","Figure 2. The same prompt through six models, with the measured cost of each call.","Five of the six printed both lines correctly. The exception is black-forest-labs/flux.2-klein-4b , which rendered the brand name correctly and misspelled the second line as SINGLE ORISION.","The most useful comparison is the cheapest model against the most expensive one on the Design Arena board. openai/gpt-image-2 billed $0.0135 and sourceful/riverflow-v2.5-pro billed $0.0654, and both produced a correct, usable packshot. Spending about five times more gave us a different look, not better text.","That does not make Riverflow overpriced. It leads the Design Arena boards, and on a harder image, an unusual composition, or a longer string of text, the gap may show. What it does mean is that neither the leaderboard nor the price column predicted the result on this prompt, and only running the prompt did.","Two smaller things came out of the same test. Your prompt length changes your bill on the per-token models, since openai/gpt-image-2 billed $0.006 on the one-line prompt earlier and $0.0135 on this longer one. And every model read “studio packshot” differently, from a hard-shadow product shot to a soft-lit one, so composition is a style choice to test.","Pick the model on whichever constraint is tightest for you. That may be cost at volume, whether the image has to contain readable words, how many reference images you need to send, or whether the reply needs text alongside the picture.","Start with openai/gpt-image-2：https://openrouter.ai/openai/gpt-image-2. It billed $0.006 on the short prompt in the cost table and $0.0135 on the longer coffee-bag prompt, because it is priced per token and your prompt is part of the bill. Budget in that range for product prompts of similar length. It is the cheapest model we measured and it rendered our text correctly for the least money.","black-forest-labs/flux.2-klein-4b：https://openrouter.ai/black-forest-labs/flux.2-klein-4b at $0.014 is the other cheap option, with one measured catch. It was the only model of the six that misspelled the label, and it had the lowest Design Arena rating of the models on both boards. Use it for volume work with no readable words in the frame. Do not use it when a product name has to come out right.","The OpenAI figures assume you leave quality unset. Moving it to high multiplied that line of our bill by 35. Klein has no quality setting, so its cost is easier to forecast and there is no setting to raise when you need a better result.","On the coffee-bag prompt the two came out at $0.0135 and $0.014, which is $135 against $140 for ten thousand images. At that point price is not deciding anything, so run your own prompts through both and let the output decide.","One operational note about Klein. Black Forest Labs publishes its weights on Hugging Face, where the repository is named black-forest-labs/FLUX.2-klein-4B：https://huggingface.co/black-forest-labs/FLUX.2-klein-4B rather than using our slug’s capitalization. You can call it through us now and move it in-house later without changing the model family you built around.","Packaging, UI mockups, slides, and posters fail when the model garbles a product name.","Five of the six models we tested handled a two-line label with no help, so pick on cost and style, then check the result against your own words rather than ours.","black-forest-labs/flux.2-flex：https://openrouter.ai/black-forest-labs/flux.2-flex rendered our label correctly, billed $0.05 for the image, and takes up to eight reference images. If you are already on FLUX.2 and the words matter, it is the FLUX.2 model to move to. If you are not committed to a family, openai/gpt-image-2 did the same job for about a quarter of the price.","Before you settle, look at recraft/recraft-v4.1-vector：https://openrouter.ai/recraft/recraft-v4.1-vector, which solves the problem differently. It returns SVG rather than pixels, and its endpoint accepts style , controls , and text_layout as passthrough parameters under provider.options . For a logo or a label that has to scale and stay editable, vector output beats a sharper raster, and $0.08 an image gets you a file you can open in a design tool.","Here, the same object has to look the same everywhere it appears. A bottle on the product page and the same bottle in a campaign image should not drift apart.","We tested this. We took the coffee bag openai/gpt-image-2 had already made, sent it back as an input_references ：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#image-to-image-reference-images image, and asked four models to keep the bag and its printed label exactly as they were while replacing the white background with a dark walnut counter under warm side light.","Figure 3. One source image, one instruction, four models, with the measured cost of each edit.","All four kept the wording and the typeface. All four also changed the bag’s shape, proportions, or lighting to some degree, and none reproduced the source bag pixel for pixel, so if every image in a campaign has to match one hero shot, compare the outputs against the source rather than reading the label alone.","To compare editing against generating fairly, we also ran the generate side using the edit instruction as the prompt, so both columns come from the same wording. That matters because two of these models bill per token, where a longer prompt is a bigger bill on its own.","black-forest-labs/flux.2-pro billed half as much again for the edit. Its edit response reported 4,096 prompt tokens where the generate response reported 34, which is the reference image being counted as input, and its endpoint lists a rate for output megapixels only, so the listed rate alone does not predict the edit price. sourceful/riverflow-v2.5-pro billed 18 percent more for the edit. On openai/gpt-image-2 and google/gemini-3-pro-image the gap was under 7 percent. These are single runs, so read the small gaps as approximate rather than as precise premiums.","Two numbers narrow the field for editing, namely how many reference images the model accepts and how it rates on the Design Arena image editing board.","black-forest-labs/flux.2-pro：https://openrouter.ai/black-forest-labs/flux.2-pro takes eight references and billed $0.03 to generate and $0.045 to edit. black-forest-labs/flux.2-max：https://openrouter.ai/black-forest-labs/flux.2-max gives you the same eight references at $0.07 per generated image.","If your brand kit runs past eight images, FLUX.2 cannot take it. The Gemini 3.x image models and Seedream 5.0 take fourteen, and the OpenAI image models take sixteen. Grok caps at three.","Sometimes the user is in a conversation and the answer should include an image and a written explanation. Nine of the 52 models on our catalog list text among their output modalities, and they are the ones that can do this. They are the Gemini image models and OpenAI’s gpt-5-image models.","Being able to return text and doing it are different things. We took four of the nine, sent each the same request through the chat completions endpoint with modalities: [\"image\", \"text\"] , asked for an image plus a two-sentence explanation, and counted how often we got both. These costs are higher than the same models’ figures in the cost table because the reply carries text, and on a per-token model you pay for both.","Every one of those calls returned at least one image. google/gemini-3-pro-image returned two images per call. The difference is the written half, and the two cheaper Gemini models mostly skipped it even though the prompt asked for it directly.","If the text is optional, google/gemini-3.1-flash-image at $0.067 works and you handle the empty case. If your product breaks without the text, google/gemini-3-pro-image at about $0.139 was the only model of the four that answered with text every time in our run.","We recommend the dedicated Image API over chat completions for new image work：https://openrouter.ai/blog/announcements/image-api, because new image models are added only to the Image API and the chat route gives you fewer parameters. openai/gpt-5-image was also the most expensive option here at $0.20 to $0.28 a call, and it returned text in two of three calls. These chat calls bill prompt and completion tokens on top of the image, so a long instruction and a long answer cost more than the per-image figure.","Generating, editing, and chatting look similar, but they are different requests. Generating and editing use the images endpoint and differ by one field. Chatting uses the chat completions endpoint instead, and only the nine text-capable models above can do it. Our Image Generation Models on OpenRouter：https://openrouter.ai/blog/tutorials/image-generation-models/ tutorial covers the full request and response contract for both endpoints.","Figure 4. Generating and editing are one endpoint with a different body. Chatting is a different endpoint.","Generating is POST /api/v1/images with a model and a prompt . You get image bytes back.","Editing is the same endpoint with input_references added, because you are changing an image you already have instead of making a new one.","Chatting sends a messages array to POST /api/v1/chat/completions . Text and images go in, and text and images can come back.","The images route is POST https://openrouter.ai/api/v1/images . The only required fields are model and prompt ."],"articleImages":[{"sourceUrl":"https://openrouter.ai/blog/images/image-generation-models-compared.png","alt":"Image Generation Models Compared: Cost, Edit, Quality","afterParagraph":0,"url":"/media/articles/cmu74ww1u0qfxrowkb6qtruwz/7b3afc695b760267.png"},{"sourceUrl":"https://openrouter.ai/blog/images/image-generation-cost-chart.png","alt":"Horizontal bar chart of the measured cost of one default-settings image on 20 image models, sorted from cheapest to most expensive. openai/gpt-image-2 is cheapest at $0.006 and google/gemini-3-pro-image is most expensive at $0.134. Each bar is labelled with its billing unit, and the per-token, per-megapixel, and per-image units are mixed throughout the range.","afterParagraph":4,"url":"/media/articles/cmu74ww1u0qfxrowkb6qtruwz/1f034c0ae78f4980.png"}],"mediaStatus":"ok","articleBodyZh":["我们通过一个 API 路由了 52 个图像生成模型：https://openrouter.ai/models?output_modalities=image。我们用同样的提示语在其中 20 个模型上运行，并记录了每次调用的计费、返回结果以及可接受的内容。","成本是最难的部分，因为这些模型的计费单位各不相同。FLUX.2 按每百万像素计价。Gemini、OpenAI 和 Microsoft MAI 按 token 计价。Grok、Recraft、Riverflow、Qwen 和 Seedream 按图像计价，费用随你请求的质量和分辨率变化。每个模型页面在其自身单位下都是正确的，因此你不能通过并排查看价格来比较它们。","我们将它们放在同一基准上。我们测试的 20 个模型都使用相同提示语和相同尺寸，然后读取每个响应的 usage.cost 字段，这就是我们收取的费用。","本页上的每个价格都来自于一次真实生成，测量日期为 2026 年 9 月 11 日。价格和模型经常变化，因此超过一个月的数字应作为参考而非预算线。你可以通过 GET /api/v1/images/models：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#via-the-image-models-api 重新获取当前状态以查看可接受的参数，或者通过 GET /api/v1/images/models/{model}/endpoints：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#per-endpoint-records 获取带单位名称的定价。","我们给这 20 个模型发送了提示语“一个玻璃瓶的影棚拍摄图，硬钥匙光，无文字。”，设置 aspect_ratio: \"1:1\"，并未使用其他设置，然后从响应中读取了 usage.cost。计费单位列是各终端定价记录中的 unit 字段。","图 1。同样的价格，按顺序排列。三个计费单位在整个范围内混合，因此单位并不能告诉你模型在其中的位置。","在根据此表制定预算前，需要了解三件事。","在 OpenAI 模型上，质量对账单的影响大于任何其他设置。我们用相同提示语在相同的 1024x1024 大小下运行 gpt-image-2，质量分别未设置、设置为低和设置为高。","未设置质量会给我们带来与低质量相同的令牌数和费用。当图像重要时，请显式设置质量，并且在未指定质量等级的情况下，将 OpenAI 模型的每图像价格视为不完整。高质量调用还花费了 123 秒返回，而默认调用只需 12 秒。","在按每百万像素计费的模型中，你请求的画面会改变费用，但变化小于像素数显示的量。我们运行了 black-forest-labs/flux.2-klein-4b：https://openrouter.ai/black-forest-labs/flux.2-klein-4b 两种分辨率。21:9 的调用返回 2400x1024，相当于 2.46 兆像素，而方形返回 1.05 兆像素。我们为方形收费 $0.014，为宽裁剪收费 $0.016。","每个模型都会在其页面和 API 的端点路由上公布费率。这些费率是计划数值，实际支付以响应为准。2026 年 9 月 11 日，sourceful/riverflow-v2.5-pro 在默认分辨率下列出了每图像 $0.13 的价格，但实际向我们收费 $0.064。black-forest-labs/flux.2-flex 列出每兆像素 $0.06 的价格，并为一兆像素图像向我们收费 $0.05。black-forest-labs/flux.2-max 列出 $0.07 并向我们精确收费 $0.07。krea/krea-2-medium 在其端点未发布定价记录，并向我们收费 $0.03。","当两者不一致时，请阅读 use.cost：https://openrouter.ai/docs/guides/overview/multimodal/image- Generation#response-format 而不是将列出的费率乘以估计值。","我们会全额收费完成的图像请求，对于失败的请求不收费。Image API 计费文档：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#billing-and-cancellation 描述了哪些结果会收费。","价格是一个维度。这些是其他维度，其中有两个会在成本之前就排除某些模型。数值来自每个端点的 supported_parameters，截至 2026 年 9 月 11 日。","参考图像是你随请求发送给模型供其参考的图像，可以是你希望编辑的图像，或是你希望匹配的风格或产品示例。可发送的图像数量上限是硬限制。Grok 支持三张，Gemini 支持十四张，因此六张图像的品牌套件会在比较其他任何内容之前排除 Grok 模型。","输出模式列是第二个轴。大多数这些模型输入文本和图像，并只返回图像字节。Gemini 图像模型和 OpenAI 的 gpt-5-image 模型也可以在同一回复中返回书面文本，这一点很重要，如果你希望模型解释它生成了什么。我们在后文测试了它们执行这一功能的可靠性。","对于任何模型，重新获取此表格请使用 GET /api/v1/images/models：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#via-the-image-models-api，该接口会发布每个模型接受的参数及其范围。supported_parameters 中不存在的参数在该端点上不被支持。","seed 是你发送的用来使结果可重复的数字。在我们目录中的所有 Gemini、Grok、OpenAI、Riverflow、Recraft 和 MAI 图像模型中都缺少此参数。FLUX.2、Seedream、Qwen 和 Krea 上存在此参数。","我们用相同的提示和 seed：424242，对 black-forest-labs/flux.2-klein-4b 运行了两次。两个响应的字节完全相同，并且都计费 $0.014。我们只测试了 Klein，所以应将其他具有 seed 功能的模型视为未测试，而不要认为它们具有相同表现。作为比较，两次调用 sourceful/riverflow-v2.5-fast 且没有 seed 参数，从同一请求体返回了两个不同的图像。","我们在每个模型页面上显示 Design Arena：https://www.designarena.ai/leaderboard/image 的评分。当我们通过基准 API：https://openrouter.ai/docs/api/api-reference/benchmarks 获取当前评分（2026 年 9 月 11 日）时，sourceful/riverflow-v2.5-pro 在图像排行榜和图像编辑排行榜上均排名最高。Microsoft 的 MAI Image 2.6 模型和 Google 的 Gemini 图像模型在图像排行榜上紧随其后，而 black-forest-labs/flux.2-klein-4b 在两个排行榜上的评分都是列出模型中最低的。","这些评分来自模型之间的直接对比投票，Design Arena 将其转换为 Elo 评分，该评分在模型胜出比赛时上升，败北时下降。","有两个前提。Gemini的评分是基于这些试镜版，而不是我们本页链接的正式发布版。Elo会在提示中告诉你投票者的偏好，而不是你的。我们系列中评分最高的模型每张图片收费0.064美元，评分最低的模型收费0.014美元。从董事会中筛选，然后根据自己的提示词做出决定。","我们向六位同价位模特发送了一个提示，并在每张图片下方标注了账单。我们要求一个哑光黑色咖啡包，上面印有两行文字。","一张工作室合照，是一个哑光黑色咖啡包，背景为纯白，硬键灯光。包正面用干净的无衬线大写字体印有“OPENROUTER ROASTERS”，下面用较小的大写字母写着“SINGLE ORIGIN”。","我们选择了易读的单词，因为它们容易评分。袋子要么写出你要求的内容，要么没有。","图2。同一提示通过六个模型，以及每个调用的测量成本。","六台中有五台正确印出了两行。例外是black-forest-labs/flux.2-klein-4b，它正确地将品牌名称表示，并将第二行拼写错误为SINGLE ORISION。","最有用的对比是Design Arena板上最便宜的型号与最贵的型号。OpenAI/GPT-Image-2售价0.0135美元，Sourceful/Riverflow-v2.5-Pro售价0.0654美元，两者都生成了正确且可用的包装照片。花费大约五倍后，画面呈现不同，文本却不变。","这并不意味着Riverflow定价过高。它领先设计竞技场的板块，在更硬的图像、不寻常的构图或较长的文字串上，差距可能会显现出来。这意味着排行榜和价格栏都没有预测这个提示的结果，只有运行提示时才预测了。","同一个测试还得出了两个小问题。你的提示词长度会影响按标记模型的账单，因为openai/gpt-image-2之前单行提示词收费0.006美元，这个较长提示词收费0.0135美元。每个模型对“studio packshot”的读法都不同，从硬阴影产品照到柔光照，所以构图是个风格选择，值得测试。","根据对您最紧迫的限制选择模型。可能是批量成本、图片中是否必须包含可读文字、需要发送的参考图片数量，或者回复是否需要文字伴随图片。","从 openai/gpt-image-2 开始：https://openrouter.ai/openai/gpt-image-2。在成本表中，短提示计费 $0.006，长咖啡袋提示计费 $0.0135，因为它按 token 定价，您的提示是计费的一部分。类似长度的产品提示预算可参考这一范围。它是我们测量的最便宜的模型，并且能以最少的成本正确渲染我们的文本。","black-forest-labs/flux.2-klein-4b：https://openrouter.ai/black-forest-labs/flux.2-klein-4b，价格 $0.014，是另一种便宜的选择，但有一个观察到的问题。它是六个模型中唯一拼写标签错误的模型，并且在两个评估板上的 Design Arena 评分都是最低的。适合生成不包含可读文字的批量作品，不适合需要正确显示产品名称的场景。","OpenAI 的计费假设您未设置质量。将质量调到高会使我们账单中的这一行乘以 35。Klein 没有质量设置，因此成本更易预测，也无法在需要更好结果时提高设置。","在咖啡袋提示下，两个模型的成本分别为 $0.0135 和 $0.014，即生成一万张图片分别为 $135 和 $140。在这种情况下，价格已经不再是决定因素，因此可以将自己的提示词分别跑在两个模型上，根据输出结果来决定。","关于 Klein 的一个操作说明。Black Forest Labs 在 Hugging Face 上发布了权重，仓库名称为 black-forest-labs/FLUX.2-klein-4B：https://huggingface.co/black-forest-labs/FLUX.2-klein-4B，而不是使用我们的 slug 的大小写。您现在可以通过我们调用它，稍后迁入内部，也不需更改您构建的模型系列。","当模型篡改产品名称时，包装、UI 模型、幻灯片和海报都会失败。","我们测试的六个模型中，有五个无需帮助就能处理两行标签，因此可以根据成本和风格选择，然后用您自己的文字检查结果，而不是依赖我们的示例。","black-forest-labs/flux.2-flex：https://openrouter.ai/black-forest-labs/flux.2-flex 正确渲染了我们的标签，图像费用为 $0.05，并且可以接受多达八张参考图像。如果你已经在使用 FLUX.2 并且文字很重要，这是必须升级到的 FLUX.2 模型。如果你没有特别偏向某个系列，openai/gpt-image-2 以大约四分之一的价格完成了相同的工作。","在你决定之前，看看 recraft/recraft-v4.1-vector：https://openrouter.ai/recraft/recraft-v4.1-vector，它以不同的方式解决了问题。它返回的是 SVG 而非像素，其端点接受 style、controls 和 text_layout 作为 provider.options 下的透传参数。对于必须可缩放且可编辑的徽标或标签，矢量输出优于更清晰的光栅图，而每张图 $0.08 可以得到一个可以在设计工具中打开的文件。","这里，同一个对象在任何地方出现时都必须看起来一致。产品页面上的瓶子和活动图片中的同一瓶子不应出现差异。","我们进行了测试。我们拿了 openai/gpt-image-2 已经制作好的咖啡袋，把它作为 input_references：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#image-to-image-reference-images 图像发送回去，并要求四个模型在保持袋子及其印刷标签完全不变的情况下，将白色背景替换为暖光下的深核桃木柜台。","图 3。一个源图像，一个指令，四个模型，每次编辑测量的费用。","四个模型都保持了文字和字体。四个模型在某种程度上也都改变了袋子的形状、比例或光照，并且没有一个像素级复刻源袋子，所以如果一整个活动中的每张图片都必须与一张主图一致，请将输出与源图像进行比较，而不仅仅是看标签。","为了公平地比较编辑与生成，我们还使用编辑指令作为提示运行了生成部分，因此两列来自相同的文字。这很重要，因为其中两个模型按 token 计费，而提示越长费用本身就越高。","black-forest-labs/flux.2-pro 对编辑的收费是生成费用的两倍半。其编辑响应报告了 4,096 个提示符标记，而生成响应只报告了 34 个标记，这是因为参考图像被计为输入，而且其端点只列出了输出百万像素的费率，所以单看列出的费率无法预测编辑价格。sourceful/riverflow-v2.5-pro 对编辑收费高出 18%。在 openai/gpt-image-2 和 google/gemini-3-pro-image 上，这一差距不到 7%。这些都是单次运行，因此应将小幅差距视为大致值，而不是精确溢价。","有两个数字可以缩小编辑模型的选择范围，即模型接受多少参考图像以及其在 Design Arena 图像编辑排行榜上的评分。","black-forest-labs/flux.2-pro：https://openrouter.ai/black-forest-labs/flux.2-pro 可接受八张参考图像，生成收费 0.03 美元，编辑收费 0.045 美元。black-forest-labs/flux.2-max：https://openrouter.ai/black-forest-labs/flux.2-max 对相同的八张参考图像每张生成图像收费 0.07 美元。","如果你的品牌素材超过八张，FLUX.2 无法处理。Gemini 3.x 图像模型和 Seedream 5.0 可接受十四张，OpenAI 图像模型可接受十六张。Grok 的上限为三张。","有时用户在对话中，希望答案包含图像和文字说明。在我们目录中的 52 个模型中，有九个在其输出模式中列出了文本，它们能够实现这一点。它们是 Gemini 图像模型和 OpenAI 的 gpt-5-image 模型。","能够返回文本和实际返回文本是不同的事情。我们选取了九个中的四个，通过聊天补全端点发送相同请求，模式设为 [\"image\", \"text\"]，要求生成图像加上两句解释，并统计我们同时获得两者的频率。这些费用高于同一模型在费用表中的数据，因为回复包含了文本，并且在按标记计费的模型上，你需要为两者付费。","所有这些调用都至少返回了一张图像。google/gemini-3-pro-image 每次调用返回两张图像。差别在于文字部分，而两个较便宜的 Gemini 模型大多数情况下直接跳过了文字，尽管提示明确要求了文字。","如果文本是可选的，google/gemini-3.1-flash-image 每次调用 $0.067 可以使用，并且可以处理为空的情况。如果您的产品在没有文本时会出问题，那么 google/gemini-3-pro-image 约 $0.139 是我们测试中四个模型中唯一每次都返回文本的模型。","我们推荐专用的 Image API 而不是用于新图像工作的聊天完成接口：https://openrouter.ai/blog/announcements/image-api，因为新图像模型只会添加到 Image API，而聊天路径可用的参数较少。openai/gpt-5-image 在这里也是最昂贵的选项，每次调用 $0.20 到 $0.28，并且在三次调用中有两次返回了文本。这些聊天调用在图像之外还会对提示和完成的 tokens 计费，因此长指令和长回答会比每张图像的费用更高。","生成、编辑和聊天看起来相似，但它们是不同的请求。生成和编辑使用 images 端点，并且只差一个字段。聊天则使用 chat completions 端点，只有上面九个支持文本的模型可以完成。我们的 OpenRouter 图像生成模型教程：https://openrouter.ai/blog/tutorials/image-generation-models/ 覆盖了两个端点的完整请求和响应协议。","图 4。生成和编辑是同一端点但请求体不同。聊天是不同的端点。","生成是 POST /api/v1/images，使用 model 和 prompt。你会获得图像字节返回。","编辑使用相同端点，但添加了 input_references，因为你是在修改已有图像而不是创建新的。","聊天将 messages 数组发送至 POST /api/v1/chat/completions。文本和图像可以输入，也可以输出文本和图像。","图像接口的请求方式是 POST https://openrouter.ai/api/v1/images。唯一必填的字段是 model 和 prompt。"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"OpenRouter 对其路由的 20 个图像生成模型使用相同提示词、相同尺寸进行实测，并记录每次调用的计费、返回结果与可接受参数。测试显示，单张图片成本介于 0.006 美元与 0.134 美元之间，差异约为 22 倍。","background":"这 20 个模型的计费单位并不一致，包括按百万像素、按 token 和按图片计费。OpenRouter 表示，测试于 2026 年 9 月 11 日完成，样本统一使用 1:1 比例和相同提示词，成本取自响应中的 usage.cost 字段。","viewpoint":"Aioga 判断：图像模型的价格不能仅依据模型页面上的标价或计费单位直接横向比较，统一提示词、尺寸和真实调用成本后，预算判断才更具可比性。","implications":"可能影响：采购或调用规划需要同时核对计费单位、质量与分辨率设置。来源显示，OpenAI 模型的质量设置会明显改变账单，因此单一价格表不足以代表所有使用情境的实际成本。","nextStep":"后续观察：需要关注模型与价格的持续变化，并在实际采购前重新获取当前模型及端点价格信息。对来源明确提示超过一个月的价格，应谨慎将其直接作为预算依据。","evidenceRefs":["title","summary","articleBody","source"],"status":"published","aiGenerated":true,"autoApproved":true,"generatedBy":"aioga-editorial:gpt-5.6-sol","reviewedBy":"aioga-editorial-review:gpt-5.6-sol","generatedAt":"2026-09-18T17:16:36.209Z","sourceHash":"330088179bda9295","review":{"approved":true,"groundedness":97,"clarity":94,"duplicationRisk":18,"blockingIssues":[],"notes":["“Aioga 判断”明确将分析性结论标注为观点，没有冒充来源事实。","“单一价格表不足以代表所有使用情境的实际成本”属于基于质量、分辨率及价格变化信息得出的合理推论，且位于“可能影响”部分，不构成阻断问题。","可选措辞优化：可将“模型页面上的标价或计费单位”改为“不同计费单位下模型页面所列的价格”，以更准确体现原文所说的不可直接并列比较。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":0,"checks":["schema","length","source-attribution","editorial-labels","inference-boundary","low-source-overlap","no-html","independent-ai-review"]}},"tags":["行业动态","OpenRouter：Announcements（RSS）"],"translations":{"zh-CN":{"title":"OpenRouter 实测 20 个图像生成模型的成本、编辑与质量","summary":"OpenRouter 对其路由的 20 个图像生成模型用相同提示词实测计费，单张图片成本在 $0.006 到 $0.134 之间，相差 22 倍。","category":"行业动态","source":"openrouter.ai","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter 实测 20 个图像生成模型的成本、编辑与质量 - Aioga AI资讯","description":"OpenRouter 对其路由的 20 个图像生成模型用相同提示词实测计费，单张图片成本在 $0.006 到 $0.134 之间，相差 22 倍。","url":"https://www.aioga.com/news/cmu74ww1u0qfxrowkb6qtruwz/","articleBody":["我们通过一个 API 路由了 52 个图像生成模型：https://openrouter.ai/models?output_modalities=image。我们用同样的提示语在其中 20 个模型上运行，并记录了每次调用的计费、返回结果以及可接受的内容。","成本是最难的部分，因为这些模型的计费单位各不相同。FLUX.2 按每百万像素计价。Gemini、OpenAI 和 Microsoft MAI 按 token 计价。Grok、Recraft、Riverflow、Qwen 和 Seedream 按图像计价，费用随你请求的质量和分辨率变化。每个模型页面在其自身单位下都是正确的，因此你不能通过并排查看价格来比较它们。","我们将它们放在同一基准上。我们测试的 20 个模型都使用相同提示语和相同尺寸，然后读取每个响应的 usage.cost 字段，这就是我们收取的费用。","本页上的每个价格都来自于一次真实生成，测量日期为 2026 年 9 月 11 日。价格和模型经常变化，因此超过一个月的数字应作为参考而非预算线。你可以通过 GET /api/v1/images/models：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#via-the-image-models-api 重新获取当前状态以查看可接受的参数，或者通过 GET /api/v1/images/models/{model}/endpoints：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#per-endpoint-records 获取带单位名称的定价。","我们给这 20 个模型发送了提示语“一个玻璃瓶的影棚拍摄图，硬钥匙光，无文字。”，设置 aspect_ratio: \"1:1\"，并未使用其他设置，然后从响应中读取了 usage.cost。计费单位列是各终端定价记录中的 unit 字段。","图 1。同样的价格，按顺序排列。三个计费单位在整个范围内混合，因此单位并不能告诉你模型在其中的位置。","在根据此表制定预算前，需要了解三件事。","在 OpenAI 模型上，质量对账单的影响大于任何其他设置。我们用相同提示语在相同的 1024x1024 大小下运行 gpt-image-2，质量分别未设置、设置为低和设置为高。","未设置质量会给我们带来与低质量相同的令牌数和费用。当图像重要时，请显式设置质量，并且在未指定质量等级的情况下，将 OpenAI 模型的每图像价格视为不完整。高质量调用还花费了 123 秒返回，而默认调用只需 12 秒。","在按每百万像素计费的模型中，你请求的画面会改变费用，但变化小于像素数显示的量。我们运行了 black-forest-labs/flux.2-klein-4b：https://openrouter.ai/black-forest-labs/flux.2-klein-4b 两种分辨率。21:9 的调用返回 2400x1024，相当于 2.46 兆像素，而方形返回 1.05 兆像素。我们为方形收费 $0.014，为宽裁剪收费 $0.016。","每个模型都会在其页面和 API 的端点路由上公布费率。这些费率是计划数值，实际支付以响应为准。2026 年 9 月 11 日，sourceful/riverflow-v2.5-pro 在默认分辨率下列出了每图像 $0.13 的价格，但实际向我们收费 $0.064。black-forest-labs/flux.2-flex 列出每兆像素 $0.06 的价格，并为一兆像素图像向我们收费 $0.05。black-forest-labs/flux.2-max 列出 $0.07 并向我们精确收费 $0.07。krea/krea-2-medium 在其端点未发布定价记录，并向我们收费 $0.03。","当两者不一致时，请阅读 use.cost：https://openrouter.ai/docs/guides/overview/multimodal/image- Generation#response-format 而不是将列出的费率乘以估计值。","我们会全额收费完成的图像请求，对于失败的请求不收费。Image API 计费文档：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#billing-and-cancellation 描述了哪些结果会收费。","价格是一个维度。这些是其他维度，其中有两个会在成本之前就排除某些模型。数值来自每个端点的 supported_parameters，截至 2026 年 9 月 11 日。","参考图像是你随请求发送给模型供其参考的图像，可以是你希望编辑的图像，或是你希望匹配的风格或产品示例。可发送的图像数量上限是硬限制。Grok 支持三张，Gemini 支持十四张，因此六张图像的品牌套件会在比较其他任何内容之前排除 Grok 模型。","输出模式列是第二个轴。大多数这些模型输入文本和图像，并只返回图像字节。Gemini 图像模型和 OpenAI 的 gpt-5-image 模型也可以在同一回复中返回书面文本，这一点很重要，如果你希望模型解释它生成了什么。我们在后文测试了它们执行这一功能的可靠性。","对于任何模型，重新获取此表格请使用 GET /api/v1/images/models：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#via-the-image-models-api，该接口会发布每个模型接受的参数及其范围。supported_parameters 中不存在的参数在该端点上不被支持。","seed 是你发送的用来使结果可重复的数字。在我们目录中的所有 Gemini、Grok、OpenAI、Riverflow、Recraft 和 MAI 图像模型中都缺少此参数。FLUX.2、Seedream、Qwen 和 Krea 上存在此参数。","我们用相同的提示和 seed：424242，对 black-forest-labs/flux.2-klein-4b 运行了两次。两个响应的字节完全相同，并且都计费 $0.014。我们只测试了 Klein，所以应将其他具有 seed 功能的模型视为未测试，而不要认为它们具有相同表现。作为比较，两次调用 sourceful/riverflow-v2.5-fast 且没有 seed 参数，从同一请求体返回了两个不同的图像。","我们在每个模型页面上显示 Design Arena：https://www.designarena.ai/leaderboard/image 的评分。当我们通过基准 API：https://openrouter.ai/docs/api/api-reference/benchmarks 获取当前评分（2026 年 9 月 11 日）时，sourceful/riverflow-v2.5-pro 在图像排行榜和图像编辑排行榜上均排名最高。Microsoft 的 MAI Image 2.6 模型和 Google 的 Gemini 图像模型在图像排行榜上紧随其后，而 black-forest-labs/flux.2-klein-4b 在两个排行榜上的评分都是列出模型中最低的。","这些评分来自模型之间的直接对比投票，Design Arena 将其转换为 Elo 评分，该评分在模型胜出比赛时上升，败北时下降。","有两个前提。Gemini的评分是基于这些试镜版，而不是我们本页链接的正式发布版。Elo会在提示中告诉你投票者的偏好，而不是你的。我们系列中评分最高的模型每张图片收费0.064美元，评分最低的模型收费0.014美元。从董事会中筛选，然后根据自己的提示词做出决定。","我们向六位同价位模特发送了一个提示，并在每张图片下方标注了账单。我们要求一个哑光黑色咖啡包，上面印有两行文字。","一张工作室合照，是一个哑光黑色咖啡包，背景为纯白，硬键灯光。包正面用干净的无衬线大写字体印有“OPENROUTER ROASTERS”，下面用较小的大写字母写着“SINGLE ORIGIN”。","我们选择了易读的单词，因为它们容易评分。袋子要么写出你要求的内容，要么没有。","图2。同一提示通过六个模型，以及每个调用的测量成本。","六台中有五台正确印出了两行。例外是black-forest-labs/flux.2-klein-4b，它正确地将品牌名称表示，并将第二行拼写错误为SINGLE ORISION。","最有用的对比是Design Arena板上最便宜的型号与最贵的型号。OpenAI/GPT-Image-2售价0.0135美元，Sourceful/Riverflow-v2.5-Pro售价0.0654美元，两者都生成了正确且可用的包装照片。花费大约五倍后，画面呈现不同，文本却不变。","这并不意味着Riverflow定价过高。它领先设计竞技场的板块，在更硬的图像、不寻常的构图或较长的文字串上，差距可能会显现出来。这意味着排行榜和价格栏都没有预测这个提示的结果，只有运行提示时才预测了。","同一个测试还得出了两个小问题。你的提示词长度会影响按标记模型的账单，因为openai/gpt-image-2之前单行提示词收费0.006美元，这个较长提示词收费0.0135美元。每个模型对“studio packshot”的读法都不同，从硬阴影产品照到柔光照，所以构图是个风格选择，值得测试。","根据对您最紧迫的限制选择模型。可能是批量成本、图片中是否必须包含可读文字、需要发送的参考图片数量，或者回复是否需要文字伴随图片。","从 openai/gpt-image-2 开始：https://openrouter.ai/openai/gpt-image-2。在成本表中，短提示计费 $0.006，长咖啡袋提示计费 $0.0135，因为它按 token 定价，您的提示是计费的一部分。类似长度的产品提示预算可参考这一范围。它是我们测量的最便宜的模型，并且能以最少的成本正确渲染我们的文本。","black-forest-labs/flux.2-klein-4b：https://openrouter.ai/black-forest-labs/flux.2-klein-4b，价格 $0.014，是另一种便宜的选择，但有一个观察到的问题。它是六个模型中唯一拼写标签错误的模型，并且在两个评估板上的 Design Arena 评分都是最低的。适合生成不包含可读文字的批量作品，不适合需要正确显示产品名称的场景。","OpenAI 的计费假设您未设置质量。将质量调到高会使我们账单中的这一行乘以 35。Klein 没有质量设置，因此成本更易预测，也无法在需要更好结果时提高设置。","在咖啡袋提示下，两个模型的成本分别为 $0.0135 和 $0.014，即生成一万张图片分别为 $135 和 $140。在这种情况下，价格已经不再是决定因素，因此可以将自己的提示词分别跑在两个模型上，根据输出结果来决定。","关于 Klein 的一个操作说明。Black Forest Labs 在 Hugging Face 上发布了权重，仓库名称为 black-forest-labs/FLUX.2-klein-4B：https://huggingface.co/black-forest-labs/FLUX.2-klein-4B，而不是使用我们的 slug 的大小写。您现在可以通过我们调用它，稍后迁入内部，也不需更改您构建的模型系列。","当模型篡改产品名称时，包装、UI 模型、幻灯片和海报都会失败。","我们测试的六个模型中，有五个无需帮助就能处理两行标签，因此可以根据成本和风格选择，然后用您自己的文字检查结果，而不是依赖我们的示例。","black-forest-labs/flux.2-flex：https://openrouter.ai/black-forest-labs/flux.2-flex 正确渲染了我们的标签，图像费用为 $0.05，并且可以接受多达八张参考图像。如果你已经在使用 FLUX.2 并且文字很重要，这是必须升级到的 FLUX.2 模型。如果你没有特别偏向某个系列，openai/gpt-image-2 以大约四分之一的价格完成了相同的工作。","在你决定之前，看看 recraft/recraft-v4.1-vector：https://openrouter.ai/recraft/recraft-v4.1-vector，它以不同的方式解决了问题。它返回的是 SVG 而非像素，其端点接受 style、controls 和 text_layout 作为 provider.options 下的透传参数。对于必须可缩放且可编辑的徽标或标签，矢量输出优于更清晰的光栅图，而每张图 $0.08 可以得到一个可以在设计工具中打开的文件。","这里，同一个对象在任何地方出现时都必须看起来一致。产品页面上的瓶子和活动图片中的同一瓶子不应出现差异。","我们进行了测试。我们拿了 openai/gpt-image-2 已经制作好的咖啡袋，把它作为 input_references：https://openrouter.ai/docs/guides/overview/multimodal/image-generation#image-to-image-reference-images 图像发送回去，并要求四个模型在保持袋子及其印刷标签完全不变的情况下，将白色背景替换为暖光下的深核桃木柜台。","图 3。一个源图像，一个指令，四个模型，每次编辑测量的费用。","四个模型都保持了文字和字体。四个模型在某种程度上也都改变了袋子的形状、比例或光照，并且没有一个像素级复刻源袋子，所以如果一整个活动中的每张图片都必须与一张主图一致，请将输出与源图像进行比较，而不仅仅是看标签。","为了公平地比较编辑与生成，我们还使用编辑指令作为提示运行了生成部分，因此两列来自相同的文字。这很重要，因为其中两个模型按 token 计费，而提示越长费用本身就越高。","black-forest-labs/flux.2-pro 对编辑的收费是生成费用的两倍半。其编辑响应报告了 4,096 个提示符标记，而生成响应只报告了 34 个标记，这是因为参考图像被计为输入，而且其端点只列出了输出百万像素的费率，所以单看列出的费率无法预测编辑价格。sourceful/riverflow-v2.5-pro 对编辑收费高出 18%。在 openai/gpt-image-2 和 google/gemini-3-pro-image 上，这一差距不到 7%。这些都是单次运行，因此应将小幅差距视为大致值，而不是精确溢价。","有两个数字可以缩小编辑模型的选择范围，即模型接受多少参考图像以及其在 Design Arena 图像编辑排行榜上的评分。","black-forest-labs/flux.2-pro：https://openrouter.ai/black-forest-labs/flux.2-pro 可接受八张参考图像，生成收费 0.03 美元，编辑收费 0.045 美元。black-forest-labs/flux.2-max：https://openrouter.ai/black-forest-labs/flux.2-max 对相同的八张参考图像每张生成图像收费 0.07 美元。","如果你的品牌素材超过八张，FLUX.2 无法处理。Gemini 3.x 图像模型和 Seedream 5.0 可接受十四张，OpenAI 图像模型可接受十六张。Grok 的上限为三张。","有时用户在对话中，希望答案包含图像和文字说明。在我们目录中的 52 个模型中，有九个在其输出模式中列出了文本，它们能够实现这一点。它们是 Gemini 图像模型和 OpenAI 的 gpt-5-image 模型。","能够返回文本和实际返回文本是不同的事情。我们选取了九个中的四个，通过聊天补全端点发送相同请求，模式设为 [\"image\", \"text\"]，要求生成图像加上两句解释，并统计我们同时获得两者的频率。这些费用高于同一模型在费用表中的数据，因为回复包含了文本，并且在按标记计费的模型上，你需要为两者付费。","所有这些调用都至少返回了一张图像。google/gemini-3-pro-image 每次调用返回两张图像。差别在于文字部分，而两个较便宜的 Gemini 模型大多数情况下直接跳过了文字，尽管提示明确要求了文字。","如果文本是可选的，google/gemini-3.1-flash-image 每次调用 $0.067 可以使用，并且可以处理为空的情况。如果您的产品在没有文本时会出问题，那么 google/gemini-3-pro-image 约 $0.139 是我们测试中四个模型中唯一每次都返回文本的模型。","我们推荐专用的 Image API 而不是用于新图像工作的聊天完成接口：https://openrouter.ai/blog/announcements/image-api，因为新图像模型只会添加到 Image API，而聊天路径可用的参数较少。openai/gpt-5-image 在这里也是最昂贵的选项，每次调用 $0.20 到 $0.28，并且在三次调用中有两次返回了文本。这些聊天调用在图像之外还会对提示和完成的 tokens 计费，因此长指令和长回答会比每张图像的费用更高。","生成、编辑和聊天看起来相似，但它们是不同的请求。生成和编辑使用 images 端点，并且只差一个字段。聊天则使用 chat completions 端点，只有上面九个支持文本的模型可以完成。我们的 OpenRouter 图像生成模型教程：https://openrouter.ai/blog/tutorials/image-generation-models/ 覆盖了两个端点的完整请求和响应协议。","图 4。生成和编辑是同一端点但请求体不同。聊天是不同的端点。","生成是 POST /api/v1/images，使用 model 和 prompt。你会获得图像字节返回。","编辑使用相同端点，但添加了 input_references，因为你是在修改已有图像而不是创建新的。","聊天将 messages 数组发送至 POST /api/v1/chat/completions。文本和图像可以输入，也可以输出文本和图像。","图像接口的请求方式是 POST https://openrouter.ai/api/v1/images。唯一必填的字段是 model 和 prompt。"]},"en":{"title":"OpenRouter actual test of the cost, editing, and quality of 20 image generation models","summary":"OpenRouter tested billing for its 20 image generation models using the same prompt, and the cost per image ranged from $0.006 to $0.134, a difference of 22 times.","category":"Industry","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter actual test of the cost, editing, and quality of 20 image generation models - Aioga AI News","description":"OpenRouter tested billing for its 20 image generation models using the same prompt, and the cost per image ranged from $0.006 to $0.134, a difference of 22 times.","url":"https://www.aioga.com/en/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:03:37.460Z"},"ja":{"title":"OpenRouter 実測 20 個の画像生成モデルのコスト、編集と品質","summary":"OpenRouter はそのルーターの 20 の画像生成モデルに対して同じプロンプトを使って実際に課金を測定したところ、1 枚あたりの画像のコストは $0.006 から $0.134 の間で、22 倍の差があった。","category":"業界動向","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter 実測 20 個の画像生成モデルのコスト、編集と品質 - Aioga AIニュース","description":"OpenRouter はそのルーターの 20 の画像生成モデルに対して同じプロンプトを使って実際に課金を測定したところ、1 枚あたりの画像のコストは $0.006 から $0.134 の間で、22 倍の差があった。","url":"https://www.aioga.com/ja/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:03:50.015Z"},"ko":{"title":"OpenRouter 실제 테스트 20개의 이미지 생성 모델의 비용, 편집 및 품질","summary":"OpenRouter는 자신의 라우터를 통해 20개의 이미지 생성 모델에 동일한 프롬프트로 실험적으로 요금을 측정했으며, 한 장의 이미지 비용은 $0.006에서 $0.134 사이로, 22배 차이가 났습니다.","category":"업계 동향","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter 실제 테스트 20개의 이미지 생성 모델의 비용, 편집 및 품질 - Aioga AI 뉴스","description":"OpenRouter는 자신의 라우터를 통해 20개의 이미지 생성 모델에 동일한 프롬프트로 실험적으로 요금을 측정했으며, 한 장의 이미지 비용은 $0.006에서 $0.134 사이로, 22배 차이가 났습니다.","url":"https://www.aioga.com/ko/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:04:46.108Z"},"es":{"title":"OpenRouter prueba los costos, la edición y la calidad de 20 modelos de generación de imágenes","summary":"OpenRouter probó la facturación de sus 20 modelos de generación de imágenes con la misma indicación; el costo por imagen varía entre $0.006 y $0.134, con una diferencia de 22 veces.","category":"Industria","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter prueba los costos, la edición y la calidad de 20 modelos de generación de imágenes - Aioga Noticias de IA","description":"OpenRouter probó la facturación de sus 20 modelos de generación de imágenes con la misma indicación; el costo por imagen varía entre $0.006 y $0.134, con una diferencia de 22 veces...","url":"https://www.aioga.com/es/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:04:40.576Z"},"fr":{"title":"OpenRouter teste en conditions réelles le coût, l'édition et la qualité de 20 modèles de génération d'images","summary":"OpenRouter a testé la facturation de ses 20 modèles de génération d'images pour le routage avec le même prompt, le coût d'une seule image se situant entre 0,006 $ et 0,134 $, soit un écart de 22 fois.","category":"Industrie","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter teste en conditions réelles le coût, l'édition et la qualité de 20 modèles de génération d'images - Aioga Actualités IA","description":"OpenRouter a testé la facturation de ses 20 modèles de génération d'images pour le routage avec le même prompt, le coût d'une seule image se situant entre 0,006 $ et 0,134 $, soit...","url":"https://www.aioga.com/fr/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:05:39.261Z"},"de":{"title":"OpenRouter testete die Kosten, Bearbeitung und Qualität von 20 Bildgenerierungsmodellen","summary":"OpenRouter testete 20 Bildgenerierungsmodelle seines Routers mit demselben Prompt und die Kosten pro Bild liegen zwischen 0,006 $ und 0,134 $, was einem Unterschied von 22-mal entspricht.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter testete die Kosten, Bearbeitung und Qualität von 20 Bildgenerierungsmodellen - Aioga KI-News","description":"OpenRouter testete 20 Bildgenerierungsmodelle seines Routers mit demselben Prompt und die Kosten pro Bild liegen zwischen 0,006 $ und 0,134 $, was einem Unterschied von 22-mal ents...","url":"https://www.aioga.com/de/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:05:37.704Z"},"pt-BR":{"title":"OpenRouter Testou o Custo, Edição e Qualidade de 20 Modelos de Geração de Imagens","summary":"O OpenRouter testou a cobrança de seus 20 modelos de geração de imagens de roteamento usando o mesmo prompt, com o custo por imagem variando de $0,006 a $0,134, uma diferença de 22 vezes.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter Testou o Custo, Edição e Qualidade de 20 Modelos de Geração de Imagens - Aioga Notícias de IA","description":"O OpenRouter testou a cobrança de seus 20 modelos de geração de imagens de roteamento usando o mesmo prompt, com o custo por imagem variando de $0,006 a $0,134, uma diferença de 22...","url":"https://www.aioga.com/pt-BR/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:06:30.348Z"},"ru":{"title":"OpenRouter фактическое тестирование стоимости, редактирования и качества 20 моделей генерации изображений","summary":"OpenRouter протестировал 20 моделей генерации изображений через его маршрутизатор с одинаковыми подсказками, стоимость одного изображения составила от $0,006 до $0,134, разница в 22 раза.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter фактическое тестирование стоимости, редактирования и качества 20 моделей генерации изображений - Aioga Новости ИИ","description":"OpenRouter протестировал 20 моделей генерации изображений через его маршрутизатор с одинаковыми подсказками, стоимость одного изображения составила от $0,006 до $0,134, разница в 2...","url":"https://www.aioga.com/ru/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:06:43.266Z"},"ar":{"title":"تجربة OpenRouter العملية لتكلفة وتحرير وجودة 20 نموذجًا لتوليد الصور","summary":"اختبرت OpenRouter 20 نموذج إنشاء الصور الخاصة بتوجيهها باستخدام نفس الكلمات الدلالية، وكانت تكلفة الصورة الواحدة تتراوح بين 0.006 دولار و0.134 دولار، بفارق 22 مرة.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"تجربة OpenRouter العملية لتكلفة وتحرير وجودة 20 نموذجًا لتوليد الصور - Aioga أخبار الذكاء الاصطناعي","description":"اختبرت OpenRouter 20 نموذج إنشاء الصور الخاصة بتوجيهها باستخدام نفس الكلمات الدلالية، وكانت تكلفة الصورة الواحدة تتراوح بين 0.006 دولار و0.134 دولار، بفارق 22 مرة.","url":"https://www.aioga.com/ar/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:07:34.790Z"},"hi":{"title":"OpenRouter ने 20 इमेज जेनरेशन मॉडल की लागत, संपादन और गुणवत्ता का वास्तविक परीक्षण किया","summary":"OpenRouter ने अपने राउटर के 20 इमेज जनरेटिंग मॉडल पर एक ही प्रॉम्प्ट का उपयोग करके वास्तविक लागत का परीक्षण किया, प्रत्येक छवि की लागत $0.006 से $0.134 के बीच थी, जो 22 गुना का अंतर दर्शाता है।","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter ने 20 इमेज जेनरेशन मॉडल की लागत, संपादन और गुणवत्ता का वास्तविक परीक्षण किया - Aioga AI समाचार","description":"OpenRouter ने अपने राउटर के 20 इमेज जनरेटिंग मॉडल पर एक ही प्रॉम्प्ट का उपयोग करके वास्तविक लागत का परीक्षण किया, प्रत्येक छवि की लागत $0.006 से $0.134 के बीच थी, जो 22 गुना का अंत...","url":"https://www.aioga.com/hi/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:07:39.245Z"},"it":{"title":"OpenRouter: test pratico dei costi, modifiche e qualità di 20 modelli di generazione di immagini","summary":"OpenRouter ha testato con lo stesso prompt 20 modelli di generazione di immagini del suo router, con un costo per immagine che va da $0,006 a $0,134, con una differenza di 22 volte.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter: test pratico dei costi, modifiche e qualità di 20 modelli di generazione di immagini - Aioga Notizie IA","description":"OpenRouter ha testato con lo stesso prompt 20 modelli di generazione di immagini del suo router, con un costo per immagine che va da $0,006 a $0,134, con una differenza di 22 volte...","url":"https://www.aioga.com/it/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:08:37.314Z"},"nl":{"title":"OpenRouter test van de kosten, bewerking en kwaliteit van 20 beeldgeneratiemodellen","summary":"OpenRouter heeft de 20 beeldgeneratiemodellen van zijn router getest met hetzelfde prompt-woord en de kosten per afbeelding liggen tussen $0,006 en $0,134, een verschil van 22 keer.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter test van de kosten, bewerking en kwaliteit van 20 beeldgeneratiemodellen - Aioga AI-nieuws","description":"OpenRouter heeft de 20 beeldgeneratiemodellen van zijn router getest met hetzelfde prompt-woord en de kosten per afbeelding liggen tussen $0,006 en $0,134, een verschil van 22 keer...","url":"https://www.aioga.com/nl/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:08:33.849Z"},"tr":{"title":"OpenRouter, 20 adet görüntü üretim modelinin maliyetini, düzenlemesini ve kalitesini test etti","summary":"OpenRouter, yönlendirdiği 20 adet görüntü oluşturma modelini aynı istem ile test ederek ücretlendirdi; tek bir resmin maliyeti $0.006 ile $0.134 arasında değişiyor, fark 22 kat.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter, 20 adet görüntü üretim modelinin maliyetini, düzenlemesini ve kalitesini test etti - Aioga AI Haberleri","description":"OpenRouter, yönlendirdiği 20 adet görüntü oluşturma modelini aynı istem ile test ederek ücretlendirdi; tek bir resmin maliyeti $0.006 ile $0.134 arasında değişiyor, fark 22 kat.","url":"https://www.aioga.com/tr/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:09:41.966Z"},"vi":{"title":"OpenRouter Thực nghiệm chi phí, chỉnh sửa và chất lượng của 20 mô hình tạo hình ảnh","summary":"OpenRouter đã thực nghiệm tính phí cho 20 mô hình tạo hình ảnh của nó bằng cùng một từ khóa, chi phí cho mỗi bức ảnh dao động từ $0,006 đến $0,134, chênh lệch 22 lần.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter Thực nghiệm chi phí, chỉnh sửa và chất lượng của 20 mô hình tạo hình ảnh - Tin tức AI Aioga","description":"OpenRouter đã thực nghiệm tính phí cho 20 mô hình tạo hình ảnh của nó bằng cùng một từ khóa, chi phí cho mỗi bức ảnh dao động từ $0,006 đến $0,134, chênh lệch 22 lần.","url":"https://www.aioga.com/vi/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:09:34.360Z"},"id":{"title":"OpenRouter Menguji Biaya, Penyuntingan, dan Kualitas dari 20 Model Generatif Gambar","summary":"OpenRouter menguji biaya 20 model generasi gambar rutenya dengan kata kunci yang sama, biaya per gambar berkisar antara $0,006 hingga $0,134, berbeda 22 kali lipat.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter Menguji Biaya, Penyuntingan, dan Kualitas dari 20 Model Generatif Gambar - Berita AI Aioga","description":"OpenRouter menguji biaya 20 model generasi gambar rutenya dengan kata kunci yang sama, biaya per gambar berkisar antara $0,006 hingga $0,134, berbeda 22 kali lipat.","url":"https://www.aioga.com/id/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:10:28.408Z"},"th":{"title":"OpenRouter ทดสอบจริงค่าใช้จ่าย การแก้ไข และคุณภาพของโมเดลสร้างภาพ 20 รุ่น","summary":"OpenRouter ทำการทดสอบการคิดค่าบริการบนโมเดลสร้างภาพ 20 แบบของเราเตอร์ด้วยคำสั่งเดียวกัน ราคาต่อภาพอยู่ระหว่าง $0.006 ถึง $0.134 แตกต่างกัน 22 เท่า","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter ทดสอบจริงค่าใช้จ่าย การแก้ไข และคุณภาพของโมเดลสร้างภาพ 20 รุ่น - ข่าว AI Aioga","description":"OpenRouter ทำการทดสอบการคิดค่าบริการบนโมเดลสร้างภาพ 20 แบบของเราเตอร์ด้วยคำสั่งเดียวกัน ราคาต่อภาพอยู่ระหว่าง $0.006 ถึง $0.134 แตกต่างกัน 22 เท่า","url":"https://www.aioga.com/th/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:10:35.328Z"},"pl":{"title":"OpenRouter testuje koszty, edycję i jakość 20 modeli generowania obrazów","summary":"OpenRouter przetestował opłaty za swoje 20 modeli generowania obrazów na podstawie tych samych promptów, koszt pojedynczego obrazu waha się od 0,006 do 0,134 USD, różnica wynosi 22 razy.","category":"行业动态","source":"OpenRouter：Announcements（RSS）","aggregationSource":"OpenRouter：Announcements（RSS）","pageTitle":"OpenRouter testuje koszty, edycję i jakość 20 modeli generowania obrazów - Aioga Wiadomości AI","description":"OpenRouter przetestował opłaty za swoje 20 modeli generowania obrazów na podstawie tych samych promptów, koszt pojedynczego obrazu waha się od 0,006 do 0,134 USD, różnica wynosi 22...","url":"https://www.aioga.com/pl/news/cmu74ww1u0qfxrowkb6qtruwz/","contentTranslated":true,"sourceHash":"e81bf69d5a83e6e3","translatedAt":"2026-09-18T17:11:29.186Z"}},"evidenceTier":"verified-news","reviewStatus":"editorial-selected","indexable":true,"editorialCover":""}}