
Claude Code terminal
Use a Sanad API key with a Claude Code-compatible endpoint for terminal-first coding sessions.
Affordable, pay-as-you-go access to the Sanad 0.1 research model, direct model access to Kimi K2.7 Code, Qwen3.8-27B, GLM 5.2, GLM 4.7 Flash, GPT OSS 120B, DeepSeek V4 Flash, and DeepSeek V4 Pro for developers and development teams. Add USD balance and use the model that fits the task in Sanad CLI, VS Code, or a compatible API workflow—without managing provider API keys.

Use a Sanad API key with a Claude Code-compatible endpoint for terminal-first coding sessions.

Install from Visual Studio Marketplace, sign in with your Sanad account, run Sanad: Select Funding, and choose from the current account model catalog.

Install from npm, run sanad login, then sanad funding, and open the interactive coding agent with the current account model catalog.
Company workspaces, member invitations, model controls, billing readiness, and usage analytics for teams adopting Sanad.
Invite members, assign admins, separate company access from personal access, and keep policy managed in one console.
Review requests, active users, model mix, client mix, and spend signals without exposing prompt bodies on the public site.
Set allowed models, compatibility-token access, monthly spend limits, and billing readiness for business adoption.
Choose the evolving Sanad 0.1 research model for Sanad's default coding experience, direct Kimi K2.7 Code for the original Kimi coding experience, Qwen3.8-27B for long-context reasoning and agentic tool use, GLM 5.2 for long-context work, GLM 4.7 Flash and DeepSeek V4 Flash for fast iterations, or GPT OSS 120B and DeepSeek V4 Pro for reasoning-heavy tasks.
Sanad's evolving research model, currently built on a Kimi K2.7 Code foundation, adds Sanad's own continuously improved research prompt and behavior for coding work.
Use Kimi K2.7 Code directly through the same Sanad balance, clients, and usage-only pricing when you want the original model experience without the Sanad 0.1 research layer.
Qwen3.8-27B combines long-context reasoning, tool use, vision, and responsive general-purpose coding support in the same Sanad clients.
GLM's official report presents strong coding, agentic, and long-context performance for software engineering workflows.
GLM 4.7 Flash provides a lower-latency option for short coding turns, explanations, and focused implementation help.
OpenAI's open-weight reasoning model gives Sanad users another option for complex debugging and agentic coding work.
DeepSeek V4 Flash gives Sanad users a responsive option for focused edits, explanations, short implementation loops, and project-aware coding help.
DeepSeek V4 Pro is available for deeper planning, debugging, code review, and longer software engineering tasks inside the same Sanad workflow.
Sanad links directly to the official GLM 5.2 report so engineers can inspect the published coding and agentic benchmark data.
Users and businesses top up USD balance and pay only for model tokens used through Sanad. No subscription, seat, workspace, or platform fee.
| Model | Input / 1M tokens | Cached input / 1M tokens | Output / 1M tokens |
|---|---|---|---|
| Sanad 0.1 | $1.05 | $0.29 | $4.10 |
| Kimi K2.7 Code | $1.05 | $0.29 | $4.10 |
| Qwen3.8-27B | $0.55 | $0.55 | $3.30 |
| GLM 5.2 | $1.50 | $0.36 | $4.50 |
| GLM 4.7 Flash | $0.16 | $0.10 | $0.50 |
| GPT OSS 120B | $0.45 | $0.10 | $0.85 |
| DeepSeek V4 Flash | $0.54 | $0.114 | $1.42 |
| DeepSeek V4 Pro | $1.42 | $0.144 | $4.06 |
Prices are in USD and apply to actual token usage. Cached-input pricing applies when cached tokens are reported for the selected model.
Self-serve, balance-based access to the current Sanad model catalog.
Add USD balance