Claude Opus 4.8

LLM Provider / Model · Long-horizon agentic execution, knowledge work and memory-backed agents

At a glance

Input price	$5
Output price	$25
Cache price	$0.50
Price tier	Premium
Context window	1M+
Max output	128K
Context tier	500K+
Speed tier	Standard
Latency	medium
Knowledge cutoff	Jan 2026
Modality	Multimodal (vision)
Model ID	claude-opus-4-8
Provider	Anthropic
Benchmark	Strongest Opus 4 series model on long-horizon agentic work
Released	May 2026
Hosting	Closed/API
Capabilities	Vision, Tool use, Adaptive thinking, Effort control (low-max), Fast mode, Streaming, Structured output, Prompt caching, Task budgets, Mid-conversation system prompts

What Claude Opus 4.8 does

Vision, Tool use, Adaptive thinking, Effort control (low-max), Fast mode, Streaming, Structured output, Prompt caching, Task budgets, Mid-conversation system prompts

Best for

Long-horizon agentic execution, knowledge work and memory-backed agents

Works well with

IDE / Editor

Integration

Context / RAG

Agent / Orchestration

Other tools

Conflicts & caveats

Trae's built-in model menu may not include Claude Opus 4.8 — Trae primarily ships with Claude/GPT/Gemini/DeepSeek.
OpenAI Codex (VS Code) is tied to OpenAI's hosted GPT-5 and Codex models — Claude Opus 4.8 is not selectable inside this integration.
Gemini CLI is locked to Google Gemini models — your LLM is Claude Opus 4.8.
Devin uses Cognition's proprietary model internally — your selected Claude Opus 4.8 will be ignored.
OpenAI Codex (Cloud) runs GPT-5.1 Codex internally — your selected Claude Opus 4.8 will be ignored.
Jules runs Gemini models internally — your selected Claude Opus 4.8 will be ignored.
Lovable uses a fixed internal model — your selected Claude Opus 4.8 will be ignored.
Replit Agent 3 uses its own model selection internally — your selected Claude Opus 4.8 will be ignored.

Build a full stack around Claude Opus 4.8 — Flowpicker shows compatibility warnings before you commit.

Open the stack planner →