# Claude Opus 5

Claude Opus 5 is an Anthropic reasoning model designed for complex software development, extended agent tasks, and professional analysis. It uses adaptive thinking to adjust reasoning effort to the task. Its published context window is 1,000,000 tokens. The advertised input rate is $5.00 per million input tokens. The advertised output rate is $25.00 per million output tokens.

- Model identifier: ` anthropic/claude-opus-5 `
- Creator: Anthropic
- Model type: Text generation
- Canonical model page: [Claude Opus 5](<https://impossibl.com/anthropic/claude-opus-5>)
- API availability: Listed as serving through Impossibl.
- Creator release date: 2026-07-24

## Capabilities and limits

- Context window: 1,000,000 tokens
- Maximum output tokens: Not published in the public catalog; this is separate from the context window.
- Inputs: text, image
- Outputs: text
- Tool calling: Not published

## API access

Use ` anthropic/claude-opus-5 ` as the model identifier through the Impossibl API.

Published API endpoints:

- ` /v1/chat/completions `

[API quickstart](<https://impossibl.com/docs/agent-quickstart>) · [Models and billing documentation](<https://impossibl.com/docs/models>)

## Published API prices

All rates are in USD. An unpublished rate is unknown, not zero. Workspace-specific pricing, taxes, balance-purchase fees, billing adjustments, and tool charges can affect the amount paid.

| Usage | Published base rate | Billing unit |
| --- | --- | --- |
| Input | $5.00 | per million input tokens |
| Output | $25.00 | per million output tokens |
| Cache read | $0.50 | per million cached input tokens |
| Cache write (default retention) | $6.25 | per million tokens |
| Cache write (one-hour retention) | $10.00 | per million tokens |
| Audio input | Not published | per million audio input tokens |

## Independent benchmarks

Source: [Artificial Analysis](<https://artificialanalysis.ai/>). Last checked: 2026-09-11T10:50:14.034Z.
Configurations are evaluated separately. API scores retain their original units. Unmeasured scores are omitted.

### Claude Opus 5 (Adaptive Reasoning, Max Effort)

Source: [Evaluation source](<https://artificialanalysis.ai/models/claude-opus-5>)

| Benchmark | Score |
| --- | --- |
| Intelligence Index | 50.70 |
| Omniscience Index | 37.07 |
| GPQA | 93.23% |
| Humanity’s Last Exam | 54.87% |
| SciCode | 56.37% |
| CritPt | 29.14% |
| Terminal-Bench v2.1 | 89.14% |
| Terminal-Bench v4.0 | 48.99% |
| AA-LCR | 79.33% |
| Multilingual LCR | 55.56% |
| Harvey LAB | 93.46% |
| GDPval Elo | 1735.09 |
| GDP.pdf All-pass | 21.60% |
| AutomationBench Partial | 56.57% |
| AnalystAgent | 53.75% |
| MMMU-Pro | 84.74% |
| Tau Banking | 42.06% |
| coding index (API) | 78 |
| agentic index (API) | 56.2 |
| intelligence index (API) | 50.7 |

### Claude Opus 5 (Adaptive Reasoning, High Effort)

Source: [Evaluation source](<https://artificialanalysis.ai/models/claude-opus-5-high>)

| Benchmark | Score |
| --- | --- |
| Intelligence Index | 48.24 |
| Omniscience Index | 33.72 |
| GPQA | 93.74% |
| Humanity’s Last Exam | 52.83% |
| SciCode | 55.44% |
| CritPt | 28.29% |
| Terminal-Bench v2.1 | 87.64% |
| Terminal-Bench v4.0 | 45.96% |
| AA-LCR | 79.00% |
| Multilingual LCR | 59.44% |
| GDPval Elo | 1628.94 |
| GDP.pdf All-pass | 19.60% |
| AutomationBench Partial | 53.56% |
| MMMU-Pro | 82.43% |
| Tau Banking | 44.74% |
| coding index (API) | 76.5 |
| agentic index (API) | 52.7 |
| intelligence index (API) | 48.2 |

### Claude Opus 5 (Adaptive Reasoning, Low Effort)

Source: [Evaluation source](<https://artificialanalysis.ai/models/claude-opus-5-low>)

| Benchmark | Score |
| --- | --- |
| Intelligence Index | 39.79 |
| Omniscience Index | 28.55 |
| GPQA | 88.89% |
| Humanity’s Last Exam | 43.42% |
| SciCode | 49.19% |
| CritPt | 23.14% |
| Terminal-Bench v2.1 | 76.40% |
| Terminal-Bench v4.0 | 26.26% |
| AA-LCR | 81.33% |
| Multilingual LCR | 53.89% |
| GDPval Elo | 1371.52 |
| GDP.pdf All-pass | 17.20% |
| AutomationBench Partial | 51.79% |
| MMMU-Pro | 79.83% |
| Tau Banking | 30.31% |
| coding index (API) | 66.9 |
| agentic index (API) | 37.5 |
| intelligence index (API) | 39.8 |

### Claude Opus 5 (Adaptive Reasoning, Medium Effort)

Source: [Evaluation source](<https://artificialanalysis.ai/models/claude-opus-5-medium>)

| Benchmark | Score |
| --- | --- |
| Intelligence Index | 45.06 |
| Omniscience Index | 31.02 |
| GPQA | 91.92% |
| Humanity’s Last Exam | 51.30% |
| SciCode | 51.50% |
| CritPt | 26.86% |
| Terminal-Bench v2.1 | 86.14% |
| Terminal-Bench v4.0 | 34.34% |
| AA-LCR | 82.00% |
| Multilingual LCR | 56.11% |
| GDPval Elo | 1524.87 |
| GDP.pdf All-pass | 20.00% |
| AutomationBench Partial | 54.34% |
| MMMU-Pro | 81.62% |
| Tau Banking | 38.56% |
| coding index (API) | 74.3 |
| agentic index (API) | 47 |
| intelligence index (API) | 45.1 |

### Claude Opus 5 (Adaptive Reasoning, Xhigh Effort)

Source: [Evaluation source](<https://artificialanalysis.ai/models/claude-opus-5-xhigh>)

| Benchmark | Score |
| --- | --- |
| Intelligence Index | 49.65 |
| Omniscience Index | 35.38 |
| GPQA | 93.74% |
| Humanity’s Last Exam | 54.40% |
| SciCode | 55.67% |
| CritPt | 27.71% |
| Terminal-Bench v2.1 | 88.01% |
| Terminal-Bench v4.0 | 46.46% |
| AA-LCR | 80.33% |
| Multilingual LCR | 58.33% |
| GDPval Elo | 1708.14 |
| GDP.pdf All-pass | 21.00% |
| AutomationBench Partial | 53.22% |
| MMMU-Pro | 83.99% |
| Tau Banking | 43.30% |
| coding index (API) | 77 |
| agentic index (API) | 55.5 |
| intelligence index (API) | 49.7 |

Full benchmark data and breakdowns: [Complete benchmark snapshot](<https://api.impossibl.com/v1/models/anthropic/claude-opus-5/benchmarks>).


## Availability and retirement

The model is listed as serving. No retirement is announced in the current public catalog.

## Frequently asked questions

### What is Claude Opus 5?

Claude Opus 5 is an Anthropic reasoning model designed for complex software development, extended agent tasks, and professional analysis. It uses adaptive thinking to adjust reasoning effort to the task.

Sources: [Claude Opus 5 model documentation](<https://platform.claude.com/docs/en/models/opus-5/overview>); [Claude Opus 5 release announcement](<https://www.anthropic.com/news/claude-opus-5>)

### How much does Claude Opus 5 cost?

Claude Opus 5 costs $5.00 per million input tokens and $25.00 per million output tokens through the Impossibl API. Cache reads cost $0.50 per million tokens; cache writes cost $6.25 per million tokens; cache writes with one-hour retention cost $10.00 per million tokens. All rates are in USD.

### What is the context length of Claude Opus 5?

Claude Opus 5 has a 1,000,000-token context window through the Impossibl API.

### What inputs and outputs does Claude Opus 5 support?

Claude Opus 5 accepts text and image as input. Claude Opus 5 returns text.

### How do I use Claude Opus 5 through an API?

Claude Opus 5 is available through the Impossibl API. Use anthropic/claude-opus-5 as the model identifier with /v1/chat/completions. The quickstart explains how to create an API key and send your first request.

### When was Claude Opus 5 released?

Anthropic released Claude Opus 5 on July 24, 2026.

Sources: [Creator release announcement](<https://www.anthropic.com/news/claude-opus-5>)

## Sources and related links

Prices, limits, capabilities, endpoints, and availability come from the public Impossibl catalog. Creator documentation supplies descriptions and release dates where available; it does not override the published API offering.

- [Public Impossibl model catalog](<https://api.impossibl.com/v1/models>)
- [Full model directory as JSON](<https://impossibl.com/models.json>)
- [Claude Opus 5: model details and pricing](<https://impossibl.com/anthropic/claude-opus-5>)
- [Model directory](<https://impossibl.com/models>)
- [Claude Opus 5 model documentation](<https://platform.claude.com/docs/en/models/opus-5/overview>)
- [Claude Opus 5 release announcement](<https://www.anthropic.com/news/claude-opus-5>)