Anthropic
Claude Opus 4.7 (Non-reasoning, High Effort)
Claude Opus 4.7 (Non-reasoning, High Effort) is an Anthropic Claude Opus profile, representing the high-capability end of the Claude lineup for difficult coding, agentic, research, and computer-use tasks. This page compares the exact benchmarked variant against other Claude and non-Claude models on capability, runtime, and price.
Introducing Claude OpusOperational Metrics
Output Speed43 tok/s
First Token1.41s
Blended Price$10.00/M
Model Metadata
Queryable facts extracted from the upstream model payload.
ReleaseApr 16, 2026
Context Windown/a
Modalitiesn/a
Strength: Coding
Rank #6 across 410 models.
53.1
Strength: Overall
Rank #12 across 500 models.
51.8
Strength: SciCode
Rank #18 across 472 models.
50.1%
Watch Area: Blended $
Rank #308 across 325 models.
$10.00/M
Watch Area: Output $
Rank #308 across 325 models.
$25.00/M
Watch Area: Input $
Rank #307 across 325 models.
$5.00/M
Strength Profile
Percentile score by analysis domain.
* Cost is inverted: lower input, output, and blended prices rank higher.
Benchmark Percentiles
Higher bars mean stronger relative placement.
All Benchmarks
| Metric | Domain | Value | Rank |
|---|---|---|---|
| Artificial Analysis Intelligence Index | overall | 51.8 | #12 |
| Artificial Analysis Coding Index | coding | 53.1 | #6 |
| GPQA | reasoning | 88.5% | #22 |
| Humanity's Last Exam | reasoning | 31.2% | #21 |
| SciCode |