快来! 免费使用 OpenAI 的 GPT-5-Codex

GPT-oss:20b

GPT-oss:20b

图文对话
Gpt-oss

GPT OSS 是OpenAI 推出的重量级开放模型,面向强推理、智能体任务以及多样化开发场景

hello codex

hello codex

最近 Codex 引入了最新的 GPT-5-Codex 模型,它在 GPT-5 基座上再训练,强化了完整项目搭建、⼤规模重构、单元测试迭代和代码评审等 Agentic Coding 能力。

不少使用 Claude Code 的用户开始尝试 Codex 纷纷给予好评,最近确实一直有 Claude Code 降智的声音。

但是如何方便有性价比地使用 OpenAI 的 GPT-5-Codex 还是有不少朋友不清楚的(如果自己有 PLus 账号可直接使用)。

老胡这边探索了一个高性价比的方案,但没有时间折腾这些,这个技术方案就给我朋友去做中转了,我叫他给我一些额度,这不第一时间就放出来给朋友们,当然,免费哈哈。

使用

怎么开始使用呢?安装客户端,打开终端,执行:

npm install -g @openai/codex  --registry=https://registry.npmmirror.com

安装成功之后开始配置:

# 一条命令搞定(若 ~/.codex 不存在会自动创建)
mkdir -p ~/.codex && tee ~/.codex/config.toml >/dev/null <<'EOF'
model_provider = "v2code"
model = "gpt-5-codex"
model_reasoning_effort = "high"
disable_response_storage = true
preferred_auth_method = "apikey"

[model_providers.v2code]
name = "v2code"
base_url = "https://api.v2code.ai/v1"
wire_api = "responses"
EOF

# 配置密钥
tee ~/.codex/auth.json >/dev/null <<EOF
{
    "OPENAI_API_KEY": "你的密钥,下面进群一起体验探讨~"
}
EOF

对于 Windows 用户:

~/.codex/config.toml文件中添加以下配置:

model_provider = "v2code"
model = "gpt-5-codex"
model_reasoning_effort = "high"
disable_response_storage = true
preferred_auth_method = "apikey"

[model_providers.v2code]
name = "v2code"
base_url = "https://code.ai2api.dev/v1"
wire_api = "responses"

~/.codex/auth.json文件中配置 API 密钥:

{
    "OPENAI_API_KEY": "你的密钥"
}

直接终端输入 codex 开始使用:

codex
codex

密钥

为了让真正需要的朋友免费体验 Codex,也顺便建立一个 AI 编程交流群,请大家加我微信,我拉你进群,免费送密钥给大家使用,用到什么程度?号池用尽为止哈哈~

一定要备注 codex 加我,不然不通过哈哈。

v: howie6879

都看到这里了,就给个关注吧👀:

喜欢我的文章,可以请你右下角顺手来一波点赞&在看&分享三连么👉

您可能感兴趣的与本文相关的镜像

GPT-oss:20b

GPT-oss:20b

图文对话
Gpt-oss

GPT OSS 是OpenAI 推出的重量级开放模型,面向强推理、智能体任务以及多样化开发场景

### OpenAI GPT-4 Model Information The GPT-4 model from OpenAI represents a significant advancement over previous models, offering enhanced performance and capabilities. This new iteration not only boosts efficiency but also addresses safety concerns more effectively than its predecessors[^2]. When utilizing the `openai` Python library or Azure OpenAI service to interact with GPT-4, determining the specific version of the model is crucial for ensuring compatibility and leveraging all available features. However, direct methods within these libraries to query the exact build number or detailed revision history are limited. Instead, developers should refer to official release notes or API documentation provided by OpenAI for precise details on versions. For users interested in deploying GPT-4 safely while maximizing utility, several strategies have been explored. These include rigorous testing protocols during development phases as well as continuous monitoring post-deployment to ensure consistent behavior across various applications. Such measures contribute towards making GPT-4 an important milestone toward creating broadly useful AI systems that prioritize security without compromising functionality. Compared to earlier iterations like those found under the GPT-3 family (e.g., Davinci), which focused primarily on enhancing text generation quality through increased complexity handling and longer content creation abilities, GPT-4 introduces additional improvements aimed at both broadening applicability areas further alongside refining existing strengths even deeper into uncharted territories where human-like intelligence becomes increasingly indistinguishable from actual cognition processes observed among people interacting naturally via written language forms such as conversations held online platforms today[^3]. ```python import openai response = openai.Completion.create( engine="gpt-4", prompt="Tell me about your latest updates.", max_tokens=150, ) print(response.choices[0].text.strip()) ```
评论
成就一亿技术人!
拼手气红包6.0元
还能输入1000个字符
 
红包 添加红包
表情包 插入表情
 条评论被折叠 查看
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值