Claude Opus 4.7
Anthropic · released Apr 16, 2026
Opus 4.7 is the next generation of Anthropic's Opus family, built for long-running, asynchronous agents. Building on the coding and agentic strengths of Opus 4.6, it delivers stronger performance on...
Specification
- Context window
- 1M
- Max output
- 128K
- Knowledge cutoff
- Not stated
- Parameters
- Undisclosed
- Licence
- Proprietary
- Serving providers
- 4
- Moderated
- Yes
- Uptime
- 100.0%
Intelligence
55.0
97th percentile
Coding
73.6
Coding Index
Agentic
46.3
Agentic Index
Output speed
77 t/s
Median across providers
Latency
1.05s
Time to first token
Cost per task
$2.25
Estimated
Benchmarks
Where the score comes from
The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.
Evaluation scores
Percentage correct · higher is better
- GPQA Diamond91.4%
- τ²-bench (Telecom)88.6%
- AA-LCR (long context)75.3%
- IFBench58.6%
- SciCode54.5%
- Terminal-Bench Hard51.5%
- Humanity's Last Exam42.3%
An evaluation missing from this list was not run for this model — it is not a zero.
View as table
| Evaluation | Score |
|---|---|
| GPQA Diamond | 91.4% |
| τ²-bench (Telecom) | 88.6% |
| AA-LCR (long context) | 75.3% |
| IFBench | 58.6% |
| SciCode | 54.5% |
| Terminal-Bench Hard | 51.5% |
| Humanity's Last Exam | 42.3% |
Against its peers
Intelligence Index · this model highlighted, nearest peers in grey
- Muse Spark 1.2Meta56.8
- GPT-5.6 TerraOpenAI56.6
- GPT-5.5OpenAI56.3
- Grok 4.5xAI55.8
- Claude Sonnet 5Anthropic55.3
- Claude Opus 4.7Anthropic55.0
- Muse Spark 1.1Meta53.2
- GPT-5.4OpenAI53.1
Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.
View as table
| Model | Intelligence |
|---|---|
| Muse Spark 1.2 | 56.8 |
| GPT-5.6 Terra | 56.6 |
| GPT-5.5 | 56.3 |
| Grok 4.5 | 55.8 |
| Claude Sonnet 5 | 55.3 |
| Claude Opus 4.7 | 55.0 |
| Muse Spark 1.1 | 53.2 |
| GPT-5.4 | 53.1 |
Percentile among all indexed models
Pricing
What it costs to run
List prices per million tokens, plus what one representative task works out to.
List price
- Input / 1M tokens
- $5
- Output / 1M tokens
- $25
- Cached input / 1M
- $0.5
- Blended 3:1
- $10
One task, estimated
$2.25
- Input tokens
- 50,000
- Output tokens
- 80,000
- Profile
- Reasoning
Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).
Serving providers
Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.
Arena
Head-to-head generation quality
Elo from pairwise judgements, broken out by the kind of thing the model was asked to build.
- Overall Elo
- 1,337
- Win rate
- 60.0%
- Strongest at
- UI components
- Tournaments
- 1,955
Elo by category
Dot position on a 1,250–1,350 scale · Elo has no meaningful zero
- 1,337
- 1,324
- 1,317
- 1,311
- 1,305
- 1,270
Agent categories (full-stack apps, mobile apps) are only scored for models tested in agent harnesses.
View as table
| Category | Elo | Win rate |
|---|---|---|
| UI components | 1337 | 60.0% |
| Game development | 1324 | 60.0% |
| Websites | 1317 | 59.0% |
| Data visualisation | 1311 | 58.2% |
| 3D scenes | 1305 | 56.7% |
| SVG | 1270 | 59.2% |