Model Details
Claude Opus 4.1
Quick answer: Claude Opus 4.1 is an incremental checkpoint update to Claude Opus 4, continuing the Anthropic pattern of versioned model improvements. It maintains the $15/$75 per million token pricing and 200K context window of the Opus 4 base, with targeted improvements to agentic reliability and instruction adherence.
What Claude Opus 4.1 Is
Claude Opus 4.1 is a versioned update in Anthropic's Opus 4 series, building on the base Opus 4 release of May 2025. Anthropic follows a numbered checkpoint system (Opus 4 → 4.1 → 4.5 → 4.6 → 4.7 → 4.8) to allow teams to track capability improvements while maintaining backward-compatible API access via dated model strings.
The 4.1 update continues the focus on long-horizon agent task reliability that defined Opus 4 at launch — complex multi-step workflows, deep codebase reasoning, and sustained autonomous operation. For teams needing the absolute ceiling of Anthropic's reasoning capability, the Opus series provides a consistent upgrade path without breaking existing integrations.
Specifications
| Field | Value |
|---|---|
| Organization | Anthropic |
| Context window | 200,000 tokens |
| License | Proprietary |
| Base model | Claude Opus 4 (May 2025) |
| Modality | Multimodal (text and vision) |
Pricing
| Input (per 1M tokens) | Output (per 1M tokens) | |
|---|---|---|
| Anthropic | $15.00 | $75.00 |
Related Models & Benchmarks
- Compare: Claude Opus 4 · Claude Opus 4.6 · Claude Sonnet 4.6
- Benchmarks: SWE-bench Verified
Specs based on Anthropic's Claude 4 model family. Last updated 2026-06-19.
Claude Opus 4.1 vs Alternatives
| Model | SWE-Bench Verified | HLE | CyberGym | License |
|---|---|---|---|---|
| Claude Opus 4.1 | 74.5% | 7.92% | 25.0% | Proprietary |
| Claude Opus 4.5 | — | — | 50.6% | Proprietary |
| Claude Opus 4.8 | — | — | — | Proprietary |
Claude Opus 4.1 vs 4.5: CyberGym jumps from 25.0% to 50.6% across generations — significant security task improvement.
Frequently Asked Questions
What is Claude Opus 4.1?
Anthropic's July 2025 model scoring 74.5% SWE-Bench Verified, 7.92% HLE. Proprietary.Scores from Benchgen evaluations. Last updated 2026-07-24.