Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis of charts and documents, complex office deliverables, and coordinating parallel subagents.
The model maintains strong instruction following and tool use across extended tasks, while remaining effective at lower effort settings for workloads that prioritize latency and token efficiency.
Providers
Different companies host the same model. OpenRouter routes your request to one of them based on the routing mode you pick: Balanced (price + speed), Nitro (fastest), or Floor (cheapest).
| $2.50 | $12.50 | $0.25 | -- | -- |
Input $2.50 / Output $12.50-- tok/s-- uptime
Pricing
The average price customers actually pay for this model, next to the prices providers post. Caching and discounts mean the price actually paid is often well below the listed one.
Weighted average
Claude Opus 5 (batch) — Price history
| $0.00 | $0.00 | 0.0% | 100.0% |
- Effective in $0.00 / Effective out $0.00Token share100.0%
Performance
Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better).
Uptime
Uptime is the percentage of the past 3 days that at least one provider was responding to requests. Availability is the percentage of time that inference was successfully served. OpenRouter continuously monitors and uses the next-best provider when one returns an error.
Not enough uptime data to display yet.
When an error occurs in an upstream provider, we can recover by routing to another healthy provider, if your request filters allow it. You can access per-provider uptime data programmatically through the Endpoints API. Learn more about our load balancing and provider routing options.
Benchmarks
Scores on standardized evaluations and this model’s own outputs on shared benchmark prompts. Higher percentages are better — and each rank shows where this model lands among the models evaluated on that source.
| Source | Benchmark | Score |
|---|---|---|
| Artificial Analysis | Claude Opus 5 (Max) Intelligence Index | 44.2 |
| Artificial Analysis | Claude Opus 5 (Max) Coding Index | 78.0 |
| Artificial Analysis | Claude Opus 5 (Max) Agentic Index | 56.5 |
| Artificial Analysis | Claude Opus 5 (Max) GPQA Diamond | 93.2% |
| Artificial Analysis | Claude Opus 5 (Max) HLE | 54.9% |
| Artificial Analysis | Claude Opus 5 (Max) AA-LCR | 79.3% |
| Artificial Analysis | Claude Opus 5 (Max) τ-Bench Banking | 42.1% |
| Artificial Analysis | Claude Opus 5 (Max) GDPval-AA | 61.5% |
| Artificial Analysis | Claude Opus 5 (Max) CritPt | 29.1% |
| Artificial Analysis | Claude Opus 5 (Max) SciCode | 56.4% |
| Artificial Analysis | Claude Opus 5 (Max) Terminal-Bench 4.0 | 49.0% |
| Artificial Analysis | Claude Opus 5 (Max) Terminal-Bench 2.1 | 89.1% |
| Artificial Analysis | Claude Opus 5 (Max) AA-Omniscience Accuracy | 60.9% |
| Artificial Analysis | Claude Opus 5 (Max) AA-Omniscience Non-Hallucination Rate | 40.6% |
| Artificial Analysis | Claude Opus 5 (Xhigh) Coding Index | 77.0 |
| Artificial Analysis | Claude Opus 5 (Xhigh) Agentic Index | 55.6 |
| Artificial Analysis | Claude Opus 5 (Xhigh) GPQA Diamond | 93.7% |
| Artificial Analysis | Claude Opus 5 (Xhigh) HLE | 54.4% |
| Artificial Analysis | Claude Opus 5 (Xhigh) AA-LCR | 80.3% |
| Artificial Analysis | Claude Opus 5 (Xhigh) τ-Bench Banking | 43.3% |
| Artificial Analysis | Claude Opus 5 (Xhigh) GDPval-AA | 59.8% |
| Artificial Analysis | Claude Opus 5 (Xhigh) CritPt | 27.7% |
| Artificial Analysis | Claude Opus 5 (Xhigh) SciCode | 55.7% |
| Artificial Analysis | Claude Opus 5 (Xhigh) Terminal-Bench 4.0 | 46.5% |
| Artificial Analysis | Claude Opus 5 (Xhigh) Terminal-Bench 2.1 | 88.0% |
| Artificial Analysis | Claude Opus 5 (Medium) Coding Index | 74.3 |
| Artificial Analysis | Claude Opus 5 (Medium) Agentic Index | 46.2 |
| Artificial Analysis | Claude Opus 5 (Medium) GPQA Diamond | 91.9% |
| Artificial Analysis | Claude Opus 5 (Medium) HLE | 51.3% |
| Artificial Analysis | Claude Opus 5 (Medium) AA-LCR | 82.0% |
| Artificial Analysis | Claude Opus 5 (Medium) τ-Bench Banking | 38.6% |
| Artificial Analysis | Claude Opus 5 (Medium) GDPval-AA | 49.9% |
| Artificial Analysis | Claude Opus 5 (Medium) CritPt | 26.9% |
| Artificial Analysis | Claude Opus 5 (Medium) SciCode | 51.5% |
| Artificial Analysis | Claude Opus 5 (Medium) Terminal-Bench 4.0 | 34.3% |
| Artificial Analysis | Claude Opus 5 (Medium) Terminal-Bench 2.1 | 86.1% |
| Artificial Analysis | Claude Opus 5 (High) Coding Index | 76.5 |
| Artificial Analysis | Claude Opus 5 (High) Agentic Index | 52.3 |
| Artificial Analysis | Claude Opus 5 (High) GPQA Diamond | 93.7% |
| Artificial Analysis | Claude Opus 5 (High) HLE | 52.8% |
| Artificial Analysis | Claude Opus 5 (High) AA-LCR | 79.0% |
| Artificial Analysis | Claude Opus 5 (High) τ-Bench Banking | 44.7% |
| Artificial Analysis | Claude Opus 5 (High) GDPval-AA | 55.2% |
| Artificial Analysis | Claude Opus 5 (High) CritPt | 28.3% |
| Artificial Analysis | Claude Opus 5 (High) SciCode | 55.4% |
| Artificial Analysis | Claude Opus 5 (High) Terminal-Bench 4.0 | 46.0% |
| Artificial Analysis | Claude Opus 5 (High) Terminal-Bench 2.1 | 87.6% |
| Artificial Analysis | Claude Opus 5 (Low) Coding Index | 66.9 |
| Artificial Analysis | Claude Opus 5 (Low) Agentic Index | 36.0 |
| Artificial Analysis | Claude Opus 5 (Low) GPQA Diamond | 88.9% |
| Artificial Analysis | Claude Opus 5 (Low) HLE | 43.4% |
| Artificial Analysis | Claude Opus 5 (Low) AA-LCR | 81.3% |
| Artificial Analysis | Claude Opus 5 (Low) τ-Bench Banking | 30.3% |
| Artificial Analysis | Claude Opus 5 (Low) GDPval-AA | 40.5% |
| Artificial Analysis | Claude Opus 5 (Low) CritPt | 23.1% |
| Artificial Analysis | Claude Opus 5 (Low) SciCode | 49.2% |
| Artificial Analysis | Claude Opus 5 (Low) Terminal-Bench 4.0 | 26.3% |
| Artificial Analysis | Claude Opus 5 (Low) Terminal-Bench 2.1 | 76.4% |
| Design Arena | Claude Opus 5 Agents Arena Agenticgamedev Elo | 1205 |
| Design Arena | Claude Opus 5 Agents Arena Androidnative Elo | 1246 |
| Design Arena | Claude Opus 5 Agents Arena Blender3d Elo | 1201 |
| Design Arena | Claude Opus 5 Agents Arena Codetovideo Elo | 1237 |
| Design Arena | Claude Opus 5 Agents Arena Full Stack Elo | 1282 |
| Design Arena | Claude Opus 5 Agents Arena Godotgamedev Elo | 1256 |
| Design Arena | Claude Opus 5 Agents Arena Mobile Apps Elo | 1348 |
| Design Arena | Claude Opus 5 Agents Arena Pdf Elo | 1235 |
| Design Arena | Claude Opus 5 Agents Arena Python-Pptxslides Elo | 1225 |
| Design Arena | Claude Opus 5 Agents Arena Spreadsheet Elo | 1229 |
| Design Arena | Claude Opus 5 Agents Arena Webapps Elo | 1250 |
| Design Arena | Claude Opus 5 Models Arena 3D Elo | 1320 |
| Design Arena | Claude Opus 5 Models Arena Asciiart Elo | 1353 |
| Design Arena | Claude Opus 5 Models Arena Code Categories Elo | 1317 |
| Design Arena | Claude Opus 5 Models Arena Data Visualization Elo | 1326 |
| Design Arena | Claude Opus 5 Models Arena Game Development Elo | 1330 |
| Design Arena | Claude Opus 5 Models Arena SVG Elo | 1306 |
| Design Arena | Claude Opus 5 Models Arena UI Component Elo | 1340 |
| Design Arena | Claude Opus 5 Models Arena Website Elo | 1304 |
| Vals AI | Claude Opus 5 BioMysteryBench | 79.3% |
| Vals AI | Claude Opus 5 Code Migration | 57.5% |
| Vals AI | Claude Opus 5 CorpFin v2 | 73.2% |
| Vals AI | Claude Opus 5 CyberBench v1.1 | 65.4% |
| Vals AI | Claude Opus 5 EMB | 73.6% |
| Vals AI | Claude Opus 5 Finance Agent (v2) | 58.6% |
| Vals AI | Claude Opus 5 GPQA Diamond | 93.4% |
| Vals AI | Claude Opus 5 Harvey's Legal Agent Benchmark v1 | 6.7% |
| Vals AI | Claude Opus 5 IOI | 84.3% |
| Vals AI | Claude Opus 5 LiveCodeBench | 89.0% |
| Vals AI | Claude Opus 5 LegalBench | 87.0% |
| Vals AI | Claude Opus 5 Legal Research Bench | 55.3% |
| Vals AI | Claude Opus 5 MedCode | 63.6% |
| Vals AI | Claude Opus 5 MedScribe | 91.0% |
| Vals AI | Claude Opus 5 MMLU Pro | 91.6% |
| Vals AI | Claude Opus 5 MMMU Pro | 89.9% |
| Vals AI | Claude Opus 5 MortgageTax | 72.1% |
| Vals AI | Claude Opus 5 MysteryMechanism | 37.4% |
| Vals AI | Claude Opus 5 ProgramBench | 3.0% |
| Vals AI | Claude Opus 5 ProofBench v1.1 | 99.0% |
| Vals AI | Claude Opus 5 Public Benefits Bench v1.1 | 76.9% |
| Vals AI | Claude Opus 5 Vals RSI Index | 33.0% |
| Vals AI | Claude Opus 5 SAGE v1 | 49.4% |
| Vals AI | Claude Opus 5 Vals Science Index | 47.9% |
| Vals AI | Claude Opus 5 SkillsBench | 60.4% |
| Vals AI | Claude Opus 5 SRE Bench | 12.2% |
| Vals AI | Claude Opus 5 SWE-bench | 97.0% |
| Vals AI | Claude Opus 5 Tax Agent Bench | 47.1% |
| Vals AI | Claude Opus 5 TaxEval v2 | 75.1% |
| Vals AI | Claude Opus 5 Terminal-Bench 4.0 | 53.5% |
| Vals AI | Claude Opus 5 Terminal-Bench-Science | 27.1% |
| Vals AI | Claude Opus 5 Vals Index | 63.7% |
| Vals AI | Claude Opus 5 Vals Multimodal Index | 73.9% |
| Vals AI | Claude Opus 5 Vibe Code Bench 1-100 | 28.5% |
| Vals AI | Claude Opus 5 Vibe Code Bench v1.1 | 88.4% |
Apps
Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.
Not enough data for apps using this model yet.
Activity
Token volume and request traffic to this model over time.
API
Drop-in code to call this model. OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. The only thing that changes between models is the model slug below.
Anthropic: Claude Opus versions
Not currently available through EU in-region routing. In-region routing docs
Documentation
Guides, cookbooks, blog posts, and tools for building with Anthropic: Claude Opus 5 (batch) on OpenRouter.
Frequently asked questions
Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis of charts and documents, complex office deliverables, and coordinating parallel subagents.
The Claude Opus 5 (batch) documentation on OpenRouter includes the Claude Opus 5 Migration Guide cookbook and the Is Jev as Accurate as Frontier Models at Classification? blog post. The quickstart covers API keys and your first request, and the API reference lists every parameter.
Claude Opus 5 (batch) costs $2.50/M input tokens and $12.50/M output tokens, with separate rates for Cache Read at $0.25/M tokens, Cache Write at $3.125/M tokens, Cache Write (1h) at $5.00/M tokens, and Web Search at $10.00/1K calls.
Claude Opus 5 (batch) has a 1,000,000 token context window. It supports up to 128,000 completion tokens.
Yes. Claude Opus 5 (batch) accepts tools and tool_choice for function calling. It also supports structured outputs via a JSON schema in response_format.
Claude Opus 5 (batch) accepts text, images, and files such as PDFs as input and returns text.
Claude Opus 5 (batch) is served by 1 provider on OpenRouter: Anthropic.
Claude Fable 5.2, Claude Haiku 5.5, Claude Sonnet 5.5 and 12 more are other text models from Anthropic.
Claude Opus 5 (batch) was released on July 24, 2026.