站点详情
正在加载站点档案和最新数据。
正在加载数据
正在从服务器获取最新数据,请稍候...
正在加载站点档案和最新数据。
正在从服务器获取最新数据,请稍候...
标准长上下文检测默认开启,会额外消耗 API 额度;检测按层推进,发现失败后停止更深层探针。
错误响应暴露上游或环境信息。
Tool-call package substitution detected (AC-1.a). A malicious middleware is rewriting package-install commands on the return path -- a code-execution-level finding. Partial credential / upstream URL / environment variable leaked in error response. The relay is exposing internal plumbing that maps onto the attacker's cr
工具回显为空、被拒绝、被截断或返回了解释文字,无法证明中间层实施了危险命令改写。
流式协议检测未能得出确定结论。
身份或延迟指纹检测存在需要复核的不确定信号。
上游通道分类探针未能完成,建议结合原始报告复核连通性或鉴权状态。
长上下文探针存在错误或边界提示,需要复核。
A: 64.83.27.22
共 66 个模型:
🟢 未检测到隐藏 Token 注入。
🟡 Test A - Verbatim: identity words (system prompt) present alongside refusal without Claude self-identification - possible partial leak, verify manually
模型身份、能力表现和协议完整性独立于安全检测统计;单次观测不等同于稳定身份结论。
上方为分组倍率,下方为充值档位。这里不展开排列组合,只保留站点当前可见的全部档位。
这里只列出该站点当前可见的全部分组与倍率,不展开组合矩阵。
| 分组 | 倍率 | 用途 |
|---|---|---|
| Media | 1 | - |
| claude-award | 0.25 | - |
| claude-premium | 0.5 | - |
| claude-stable | 0.315 | - |
| claude-tmp-cheap | 0.06 | - |
| deepseek-award | 0.35 | - |
| deepseek-stable | 0.6 | - |
| gemini-premium | 0.5 | - |
| gemini-stable | 0.15 | - |
| glm-award | 0.35 | - |
| glm-stable | 0.5 | - |
| grok-award | 0.01 | - |
| grok-stable | 0.11 | - |
| kimi-award | 0.35 | - |
| kimi-stable | 0.65 | - |
| openai-award | 0.03 | - |
| openai-enterprise | 0.5 | - |
| openai-premium | 0.13 | - |
| openai-stable | 0.08 | - |
| qwen-stable | 0.8 | - |
这里只列出该站点全部充值档位,便于核对月卡、周卡、日卡和按量额度。
| 档位 | 类型 | 实付金额 | 美元额度 | 有效期 / 说明 | 充值位置 |
|---|---|---|---|---|---|
| US$99.99 · Modelflare 28-Day Subscription | 日卡 | ¥99.99 | $110.00 | 28 day subscription | public /api/subscription/plans |
| US$299.99 · Modelflare 28-Day Subscription | 日卡 | ¥299.99 | $340.00 | 28 day subscription | public /api/subscription/plans |
| US$499.99 · Modelflare 28-Day Subscription | 日卡 | ¥499.99 | $580.00 | 28 day subscription | public /api/subscription/plans |
| US$999.99 · Modelflare 28-Day Subscription | 日卡 | ¥999.99 | $1180.00 | 28 day subscription | public /api/subscription/plans |
| US$1,499.99 · Modelflare 28-Day Subscription | 日卡 | ¥1499.99 | $1800.00 | 28 day subscription | public /api/subscription/plans |
按公开页面或登录态 probe 抓取并保留最新 5 条。
🎉 Gemini 4 Neon, Hydrogen, and Carbon are live | Neon scores within 1.1 points of Opus 5.5 on FrontierCode 2.0
Gemini 4 Neon, Hydrogen, and Carbon are now available on Modelflare. Argon, already available, also joins gemini-premium. Neon is suited to complex coding and multi-step agent workflows, Hydrogen to everyday development and high-volume tasks, and Carbon to deep reasoning and large codebase analysis. Google lists a shared input and output budget of about 1.05 million tokens for all four models. Google’s Gemini 4 guide
gemini-premium bills at 50% of Google’s standard list prices. Prices per million tokens are:
gemini-4-neon $1.50 input, $0.075 cached input, $7.50 output.gemini-4-hydrogen $0.75 input, $0.0375 cached input, $3.75 output.gemini-4-carbon and gemini-4-argon $2.00 input, $0.10 cached input, $10.00 output.Benchmark results reported by Google
high thinking, Neon scored 46.2% on FrontierCode 2.0. The same report lists Claude Opus 5.5 at 47.3% and GPT-6.2 Sol at 43.8%. Google also reports 32% lower cost per task than Carbon on the Artificial Analysis Intelligence Index at Neon’s high setting. Evaluation methodology and results · Google’s launch announcementhigh setting. Google’s launch announcementThese results were published or compiled by Google. Modelflare has not independently reproduced them. Select gemini-premium as your API key’s group to use these models through the Chat Completions API. This release also adds gemini-3-flash-preview. View the model catalog
🎉 GPT-6.2 Astra and Sol are live — Astra 20% cheaper, Sol close to Astra on coding at about one-fifth the cost
OpenAI released gpt-6.2-astra and gpt-6.2-sol on 17 November 2026, and both are now available on Modelflare. gpt-6.2-astra is built for the hardest projects. OpenAI says it sets new highs on professional-work and scientific-research evaluations at less than a third of gpt-6-astra's cost per task. gpt-6.2-sol is an efficient, versatile model for everyday work. gpt-6-astra and gpt-6.1-sol stay available.
Both models have a 1.05M-token context window, with up to 922K input tokens and 128K output tokens. They accept text and image input and return text. Chat Completions and the Responses API are both supported. reasoning.effort does not accept none.
Official prices per 1M tokens:
gpt-6.2-astra: $8 input, $0.40 cached input, $10 cache writes, $40 output. Input and output are 20% below gpt-6-astra, and cached input falls from $1 to $0.40.gpt-6.2-sol: $2 input, $0.10 cached input, $2.50 cache writes, $10 output, the same as gpt-6.1-sol.When a request has more than 272K input tokens, the whole request is billed at double the input and cache prices and 1.5 times the output price.
OpenAI-reported results:
gpt-6.2-astra at xhigh effort scores 5.2 points above gpt-6-astra at max effort, at less than half the cost per task.Service fully restored | November 11–14: 11.11% friend invitation reward, US$11.11 subscription bonus
Modelflare is fully back in normal operation. You can sign in, call the API and top up as usual. We sincerely apologize for the disruption the earlier incident caused.
Balances
Please sign in with the email address you used for your earlier top-ups, or register again with that email address. Most previous balances have already been returned to the account under the same email address. If you have questions about your balance, or it does not match what you had before, please send us your top-up email address and proof of payment. We will verify and restore it as soon as we can.
API keys and invitation records
Previous API keys need to be created again and updated in your application configuration.
If you took part in friend invitations before, or are one of our affiliate partners, and your records were not restored automatically, please apply again and we will help you through it.
Limited-time offers: November 11–14 (Beijing time)
We understand that some of you may still have concerns about continuing to use the service and topping up. We are still here, and we will restore every balance that is owed. If you run into any problem, including balance checks or applying again for friend invitations or the affiliate program, please . We will do our best to resolve it for every user.
🎉 Claude Fable 5.2 is live — ahead on 10 of 12 published benchmarks, 20% cheaper per token than Fable 5.1
claude-fable-5-2 is now available on Modelflare. It is Anthropic's strongest model open to all API customers, built for hard reasoning and long-running agent work. The context window is 1M tokens, with up to 128K output tokens. The Messages API and Chat Completions are both supported. claude-fable-5-1 stays available.
Official prices per 1M tokens are $8 input and $40 output, then the group multiplier applies: claude-premium $4 / $20, claude-stable $2.52 / $12.60, claude-award $2 / $10.
claude-premium and claude-stable are more reliable and recommended for production. claude-award is the cheapest, but requests may occasionally fail for now; we are working on it.
Anthropic-reported results: 49.1% on FrontierCode 2.0 (Fable 5.1 32.4%, Opus 5.5 45.5%, GPT-6 Astra 42.8%) and 41.3% on Remote Labor Index (Fable 5.1 17.9%). At high effort it matches Opus 5.5 at xhigh on FrontierCode at about half the cost per task.
These are Anthropic's own evaluations; Modelflare has not reproduced them. Details: our deep dive, pricing page.
🎁 AnyModelBot Credits Are Here | 10% Extra for Subscribers, $1 for New Users
Modelflare now offers AnyModelBot credits, usable only in AnyModelBot:
Sign in to AnyModelBot with your Modelflare account to use them. Requests made in AnyModelBot use these credits first, then your subscription credits and account balance.
展示该站点在三个核心时间窗口下的正式排名指标。综合排名榜单默认展示全时段数据。
暂无正式排名数据。
暂无正式排名数据。
暂无正式排名数据。
展示 GitHub 用户对该站点的公开评分与留言。
暂无公开评价。
gpt-6.2-sol comes close to gpt-6.2-astra at about one-fifth the cost.These are OpenAI's own evaluations; Modelflare has not reproduced them. Details: launch page. Modelflare prices: GPT-6.2 Astra pricing, GPT-6.2 Sol pricing.
The Modelflare Team