# Claude Code

Canonical page: https://agentsky.dev/agents/claude-code

Claude's coding agent, zero model cost with your subscription.

Claude Code is Anthropic's official agentic coding assistant, built to write, edit, test, and commit code in real codebases using Claude models. On AgentSky it runs in a persistent cloud computer with file-system access, tools, and optional channels — reachable via API without a local install.

## Facts

- **Agent type:** `claude_code`
- **Default model:** Claude Opus 5.5 (`claude-opus-5-5`)
- **Start in the browser:** https://agentsky.dev/playground/enter?agentType=claude_code&llm=claude-opus-5-5&source=agent-md
- **API reference:** https://agentsky.dev/docs
- **Billing:** Running sessions are billed by compute time and model tokens; parked agents incur no compute charge.

## Models it runs on

USD per 1M tokens.

| Model | Id | Input | Output |
|---|---|---|---|
| Claude Opus 5.5 (default) | `claude-opus-5-5` | $2.80 | $14.00 |
| Claude Opus 5 | `claude-opus-5` | $3.50 | $17.50 |
| Claude Fable 5.1 | `claude-fable-5-1` | $7.00 | $35.00 |
| Claude Sonnet 5.5 | `claude-sonnet-5-5` | $1.40 | $7.00 |
| Claude Sonnet 5 | `claude-sonnet-5` | $1.40 | $7.00 |
| Claude Haiku 5.5 | `claude-haiku-5-5` | $0.07 | $0.35 |
| Claude Haiku 4.5 | `claude-haiku-4-5` | $0.70 | $3.50 |
| GPT-6.1 Sol | `gpt-6.1-sol` | $1.40 | $7.00 |
| GPT-6 Sol | `gpt-6-sol` | $1.40 | $7.00 |
| GPT-6 Luna | `gpt-6-luna` | $0.07 | $0.35 |
| GPT-5.6 Sol | `gpt-5.6-sol` | $3.50 | $21.00 |
| GPT-5.6 Terra | `gpt-5.6-terra` | $1.40 | $8.40 |
| GPT-5.6 Luna | `gpt-5.6-luna` | $0.14 | $0.84 |
| GPT-6 Astra | `gpt-6-astra` | $7.00 | $35.00 |
| DeepSeek V4 Pro | `deepseek-v4-pro` | $0.924 | $2.772 |
| DeepSeek V4.1 Flash | `deepseek-flash` | $0.21 | $0.84 |
| Kimi K3.1 Pro | `kimi-k3.1-pro` | $2.20 | $11.00 |
| Kimi K3 | `kimi-k3` | $3.30 | $16.50 |
| Kimi K2.7 Code | `kimi-k2.7-code` | $1.045 | $4.40 |
| GLM-5.3 | `glm-5.3` | $1.54 | $4.84 |
| GLM-5.3-Flash | `glm-5.3-flash` | $0.165 | $0.55 |
| Grok 4.7 | `grok-4.7` | $2.20 | $6.60 |
| Grok 4.6 | `grok-4.6` | $2.20 | $6.60 |
| Gemini 4 Argon | `gemini-4-argon` | $4.40 | $22.00 |
| Gemini 3.8 Flash | `gemini-3.8-flash` | $1.65 | $8.25 |
| Gemini 3.5 Flash-Lite | `gemini-3.5-flash-lite` | $0.33 | $2.75 |
| Muse Spark 1.3 | `muse-spark-1.3` | $1.375 | $4.675 |
| Muse Spark 1.2 | `muse-spark-1.2` | $1.375 | $4.675 |
| Muse Spark 1.1 | `muse-spark-1.1` | $1.375 | $4.675 |
| Muse Wonder 1.0 | `muse-wonder-1.0` | $4.40 | $22.00 |
| MiniMax-M3 | `minimax-m3` | $0.33 | $1.32 |
| MiniMax-M3.1-Flash | `minimax-m3.1-flash` | $0.165 | $0.66 |

## What it can do

- **Write & edit code:** Reads and modifies files in a persistent cloud workspace across the full project.
- **Run your tests:** Executes the test suite and iterates until the run is green.
- **Commit to git:** Stages, commits, and pushes — leaving a clean, auditable history.
- **Run overnight:** Keeps working while you sleep; the cloud computer never shuts down mid-task.

## Benchmarks

Best pair: Claude Code + Claude Fable 5 — #1 of 8 ranked pairs, Elo 1,201, 97.5% completed across 4,240 tasks (2026-05-06–2026-08-06). Full board: https://agentsky.dev/benchmarks

## Questions

### How do I run Claude Code in the cloud on AgentSky without installing it locally on my machine?

Sign up for AgentSky and start a Claude Code agent via the API or web UI. No local install is needed: AgentSky provisions a persistent cloud computer with Claude Code and your chosen Claude model. Pass your task and model in the same API call.

### Which Claude models can I use with Claude Code on AgentSky, and why are GPT or DeepSeek models not supported?

Claude Code on AgentSky supports Claude Opus 5.5, Claude Opus 5, Claude Fable 5.1, Claude Sonnet 5.5, Claude Sonnet 5, Claude Haiku 5.5, Claude Haiku 4.5, GPT-6.1 Sol, GPT-6 Sol, GPT-6 Luna, GPT-5.6 Sol, GPT-5.6 Terra, GPT-5.6 Luna, GPT-6 Astra, DeepSeek V4 Pro, DeepSeek V4.1 Flash, Kimi K3.1 Pro, Kimi K3, Kimi K2.7 Code, GLM-5.3, GLM-5.3-Flash, Grok 4.7, Grok 4.6, Gemini 4 Argon, Gemini 3.8 Flash, Gemini 3.5 Flash-Lite, Muse Spark 1.3, Muse Spark 1.2, Muse Spark 1.1, Muse Wonder 1.0, MiniMax-M3, and MiniMax-M3.1-Flash. GPT, DeepSeek, Gemini, and Kimi models are not compatible with the Claude Code client. Claude Fable 5, Claude Sonnet 4.6, and Claude Opus 4.8 keep running on existing agents but cannot be chosen for new ones.

### What does running Claude Code on AgentSky cost, and is there a way to use my existing Claude subscription to reduce costs?

Connect a Claude Pro or Max plan via `sky auth connect` and model usage is $0 on eligible runs. Without a connected subscription, Claude Code with Claude Fable 5 ran 4,240 Playground tasks at a median cost of $4.86 per task and a 90th-percentile cost of $16.54. Current rates are on the pricing page.

### My Claude Code agent on AgentSky times out or loses context when working on a large codebase — how do I fix this?

Context overflow is the most common cause. Claude Code accumulates file content and conversation history per turn; large codebases with many open files push runs past the model's context window. Scope each task to a bounded directory or module, and use Claude Fable 5.1, which is built for the longest-horizon work of the available models. Check the task log for the exact message where truncation occurred.

### If my Claude Code agent is actively editing files and the AgentSky cloud machine restarts, will it lose the code changes it made?

The workspace and files persist across ordinary restarts. If the underlying machine is replaced, AgentSky restores the latest durable recovery point, which typically includes the working directory and recent file edits. Work in progress at the exact moment of a hard replacement may replay from the last checkpoint. Committing intermediate results with git adds a durable recovery layer on top of AgentSky's managed checkpoints.

### How do I connect my existing Claude Pro or Max subscription to AgentSky so I do not pay per-token for Claude Code model usage?

Run `sky auth connect` and complete the Anthropic OAuth flow to link your Claude account. Once connected, Claude Code uses your subscription as the model provider and eligible model usage costs $0 instead of per-token billing. The BYO connection covers Claude Code with Claude models; Codex with ChatGPT uses the same `sky auth connect` flow with an OpenAI account. Compute, tools, and connector costs remain separate from model usage.