Claude Opus 5
ReleasedModel Claude Opus 5 is a model by Anthropic, released on 24 Jul 2026 (2 months ago).
- Where can it be used?
- Evidenced as usable in: Claude, Claude Code, Claude Enterprise, Claude Team, GitHub Copilot, Microsoft 365 Copilot (verified for Cowork only)For 8 more recorded products there is no statement; that does not mean “not available”.Available in the provider’s API
- Did we test it?
- No lab run published yet
- Last change
- 2 months ago
- Evidence
- 7 sources from 4 sitesRegistry entry:
checked 3 Oct 2026
What backs this up?
Three kinds of evidence, kept apart. None of them proves AGI on its own; where one is missing, the page says so.
- In real productsEvidenced as usable in: Claude, Claude Code, Claude Enterprise, Claude Team, GitHub Copilot, Microsoft 365 Copilot (verified for Cowork only)
- agi.jetzt PrüfstandNo published run yet.
- Public benchmarks23 reported results on 18 benchmarks — each with its source
A model and the product around it are different things: a result for one does not carry over to the other.
Public benchmark results
Results reported by benchmark maintainers or providers, each with its source. They are not averaged and not compared across benchmarks; a provider result is self-reported. Some sources are data files of the benchmark’s own site rather than pages; they are marked as machine-readable.
- Claude Opus 527.3 %Chartography · v1 · accuracy (expert-range graded), no toolssource as of 03/10/2026 · benchmark maintainer · mode not stated · judge: google/gemini-3.5-flash · tools: no · setting: Inspect AI (surge-ai/chartography) · attempts: 10 · web: no · reasoning: Adaptive thinking, max effort · computer use: no · reported as „Claude Opus 5 (Adaptive/Max)“Source: surgehq.ai — Chartography, Claude Opus 5
- Claude Opus 524 %GDP.pdf · v1 · pass rate, all criteria (strict)source as of 03/10/2026 · benchmark maintainer · mode not stated · judge: google/gemini-3.5-flash · tools: no · setting: Inspect AI (surge-ai/gdp-pdf) · attempts: 5 · web: no · reasoning: Adaptive thinking, max effort · computer use: no · reported as „Claude Opus 5 (Adaptive/Max)“Source: surgehq.ai — GDP.pdf, Claude Opus 5
- Claude Opus 553.94 %Terminal-Bench · 4.0 · resolution rate (mean of 5 runs)source as of 03/10/2026 · benchmark maintainer · model + harness · Claude Code · tools: yes · attempts: 5 · pass@k: 1 · reasoning: xhigh · reported as „Opus 5“Source: www.tbench.ai — Terminal-Bench, Claude Opus 5
- Claude Opus 5$11,181.87Vending-Bench 2 · 2 · final bank balancesource as of 03/10/2026 · benchmark maintainer · model + harness · Andon Labs Vending-Bench 2 agent environment (email, web search, inventory tools, notes/reminders) · tools: yes · budget: ~1 simulated year; 3000-6000 messages, 60-100M output tokens per run · attempts: 6 · web: yesMachine-readable source: andonlabs.com · data bundle of the website, not linked
- Claude Opus 590.42 %ARC-AGI-2 · semi-private · accuracy (task pass rate)source as of 02/10/2026 · benchmark maintainer · mode not stated · reasoning: max · reported as „Claude Opus 5 (Max)“Machine-readable source: arcprize.org · JSON — ARC-AGI-2, Claude Opus 5
- Claude Opus 530.16 %ARC-AGI-3 · semi-private · scoresource as of 02/10/2026 · benchmark maintainer · mode not stated · reasoning: high (only effort tested on v3) · reported as „Claude Opus 5 (High)“Machine-readable source: arcprize.org · JSON — ARC-AGI-3, Claude Opus 5
- Claude Opus 553.4 %FrontierCode · 1.1-main · overall score (mergeability)source as of 29/09/2026 · benchmark maintainer · mode not stated · judge: model-graded rubric checks (grader not named on leaderboard) · tools: yes · web: yesSource: cognition.com — FrontierCode, Claude Opus 5
- Claude Opus 549.7 %BenchCAD Vision2Code · v1 · voxel IoU, no toolssource as of 22/09/2026 · self-reported by the provider · mode not stated · tools: no · setting: Anthropic internal; random 1,000-file subset of 17,900 Vision2Code files; 256x256 px views (corrected) · attempts: 5 · reasoning: adaptive thinking, max effortSource: www-cdn.anthropic.com · PDF — BenchCAD Vision2Code, Claude Opus 5
- Claude Opus 589.9 %BenchCAD Vision2Code · v1 · voxel IoU, with toolssource as of 22/09/2026 · self-reported by the provider · mode not stated · tools: yes · setting: Anthropic internal; random 1,000-file subset of 17,900 Vision2Code files; 256x256 px views (corrected) · attempts: 5 · reasoning: adaptive thinking, max effortSource: www-cdn.anthropic.com · PDF — BenchCAD Vision2Code, Claude Opus 5
- Claude Opus 583.4 %Chartography · v1 · accuracy (expert-range graded), with toolssource as of 22/09/2026 · self-reported by the provider · mode not stated · judge: Gemini 3.5 Flash · tools: yes · setting: Anthropic internal implementation of Chartography · attempts: 5 · web: no · reasoning: adaptive thinking, max effort · computer use: noSource: www-cdn.anthropic.com · PDF — Chartography, Claude Opus 5
- Claude Opus 573.65 %DeepSWE · 1.1 · pass@1 (mean of 4 runs)source as of 22/09/2026 · benchmark maintainer · model + harness · mini-swe-agent · tools: yes · attempts: 4 · pass@k: 1 · reasoning: max · reported as „claude-opus-5“Source: deepswe.datacurve.ai — DeepSWE, Claude Opus 5
- Claude Opus 540.5 %HLE-Diamond · 2026-09-22 · accuracysource as of 22/09/2026 · benchmark maintainer · model alone · web: no · reasoning: maxSource: lastexam.ai — HLE-Diamond, Claude Opus 5
Show 11 more
- Claude Opus 568.9 %HLE-Diamond · 2026-09-22 · accuracy · toolssource as of 22/09/2026 · benchmark maintainer · model + harness · Claude Code 2.1.278 · tools: yes · web: yes · reasoning: highSource: lastexam.ai — HLE-Diamond, Claude Opus 5
- Claude Opus 559.4 %SWE-bench Multimodal · v1 · resolve ratesource as of 22/09/2026 · self-reported by the provider · mode not stated · tools: yes · attempts: 5 · reasoning: adaptive thinking, max effortSource: www-cdn.anthropic.com · PDF — SWE-bench Multimodal, Claude Opus 5
- Claude Opus 544.33 %OSWorld · 2.1 · binary accuracy (task fully completed)source as of 17/09/2026 · benchmark maintainer · mode not stated · tools: yes · budget: 500 steps · subset: full set · reasoning: max · computer use: yesMachine-readable source: osworld-v2.xlang.ai · JSON — OSWorld, Claude Opus 5
- Claude Opus 570.19 %OSWorld · 2.0-v2026.08.08 · partial scoresource as of 17/09/2026 · benchmark maintainer · mode not stated · tools: yes · budget: 500 steps · subset: offline set · reasoning: max · computer use: yesMachine-readable source: osworld-v2.xlang.ai · JSON — OSWorld, Claude Opus 5
- Claude Opus 577.67 %OSWorld · 2.1 · partial score (mean checkpoint credit)source as of 17/09/2026 · benchmark maintainer · mode not stated · tools: yes · budget: 500 steps · subset: full set · reasoning: max · computer use: yesMachine-readable source: osworld-v2.xlang.ai · JSON — OSWorld, Claude Opus 5
- Claude Opus 548.71 %τ³-bench banking (pass@1) · tau2-bench-1.0.1 · pass@1evaluated 03/08/2026 · benchmark maintainer · model + harness · τ²-bench harness (Sierra, v1.0.1) · tools: yes · attempts: 4 · pass@k: 1 · reasoning: maxMachine-readable source: raw.githubusercontent.com · JSON — τ³-bench banking (pass@1), Claude Opus 5
- Claude Opus 531.96 %τ³-bench banking (pass^4 consistency) · tau2-bench-1.0.1 · pass^4 (all 4 runs succeed)evaluated 03/08/2026 · benchmark maintainer · model + harness · τ²-bench harness (Sierra, v1.0.1) · tools: yes · attempts: 4 · pass@k: pass^4 · reasoning: maxMachine-readable source: raw.githubusercontent.com · JSON — τ³-bench banking (pass^4 consistency), Claude Opus 5
- Claude Opus 50.49AA-Omniscience (net score) · public-split · net score (correct − incorrect)source as of 24/07/2026 · self-reported by the provider · model alone · web: noSource: www-cdn.anthropic.com · PDF — AA-Omniscience (net score), Claude Opus 5
- Claude Opus 573.17 %FrontierMath Tier 4 · v2 · accuracy (mean score)evaluated 24/07/2026 · benchmark maintainer · mode not stated · tools: yes · budget: 1,000,000 tokens hard limit (forced submission at 660,000) · setting: Epoch AI FrontierMath implementation (Inspect): python tool + submit_answer tool · web: no · reasoning: max · reported as „claude-opus-5_max“Machine-readable source: epoch.ai · ZIP download — FrontierMath Tier 4, Claude Opus 5
- Claude Opus 585.61 %FrontierMath Tiers 1–3 · v2 · accuracy (mean score)evaluated 24/07/2026 · benchmark maintainer · mode not stated · tools: yes · budget: 1,000,000 tokens hard limit (forced submission at 660,000) · setting: Epoch AI FrontierMath implementation (Inspect): python tool + submit_answer tool · web: no · reasoning: max · reported as „claude-opus-5_max“Machine-readable source: epoch.ai · ZIP download — FrontierMath Tiers 1–3, Claude Opus 5
- Claude Opus 580.6 %SimpleBench · avg5 · accuracy (average of 5 runs)evaluated 24/07/2026 · benchmark maintainer · model alone · pass@k: AVG@5 · temperature: 0.7 (top-p 0.95)Machine-readable source: simple-bench.com · JavaScript file — SimpleBench, Claude Opus 5
Where can it be used?
Only recorded availability statements are shown. A missing plan or surface means unknown, not unavailable. Where it applies only to a plan or a region, the row says so. “Rolling out” is not “available”. “Rollout reported” means a rollout was reported, but nothing has supported it for more than 90 days.
| Product | Status | since | Checked | Sources |
|---|---|---|---|---|
| ClaudeApp | Limited availability | not documented | checked 11 Oct 2026 | 1support.claude.com |
| ClaudeApp · Max 20x | Limited availability | not documented | ||
| ClaudeApp · Max 5x | Limited availability | not documented | ||
| ClaudeApp · Pro | Limited availability | not documented | ||
| Claude CodeCode | Limited availability | not documented | checked 11 Oct 2026 | 1support.claude.com |
| Claude EnterpriseWork · Enterprise | Limited availability | not documented | ||
| Claude TeamWork · Team | Limited availability | not documented | ||
| GitHub CopilotCopilot · Business | Limited availability | not documented | checked 11 Oct 2026 | 1docs.github.com |
| GitHub CopilotCopilot · Enterprise | Limited availability | not documented | checked 11 Oct 2026 | 1docs.github.com |
| GitHub CopilotCopilot · Max | Limited availability | not documented | checked 11 Oct 2026 | 1docs.github.com |
| GitHub CopilotCopilot · Pro Plus | Limited availability | not documented | checked 11 Oct 2026 | 1docs.github.com |
| GitHub Copilot | Limited availability | not documented | checked 11 Oct 2026 | 1docs.github.com |
| Microsoft 365 CopilotCowork | Limited availability | not documented | checked 11 Oct 2026 | 1learn.microsoft.com |
Where a start is “not documented”, a source check found the state in place; the source names no start date.
No statement on record for 8 further products: ChatGPT, ChatGPT Work, Codex, Gemini, Gemini for Google Workspace, Grok, Langdock, Le Chat Enterprise. No statement is not the same as not available.
Developer APIs
Developer access — not a statement about any work product.
| Product | Status | since | Checked | Sources |
|---|---|---|---|---|
| Claude APIAPI | Available | not documented | checked 3 Oct 2026 | 1platform.claude.com |
What did we test?
No lab run of Claude Opus 5 is published yet.
Nothing is shown until a real run exists. See what the lab tests:
Lifecycle and history
The dates the provider confirmed, and every recorded period of availability. A current period shows its checked status; one without a source check is marked not verified.
Announced—not recorded
Released
Deprecated—not recorded
Retired—not recorded
What changed?
Sourced changes concerning this model, newest first.
- 2 months agoadded to the record on 30 Sept 2026
1platform.claude.com
Evidence
Every source document behind the facts on this page, by domain.
| Domain | Docs | Documents |
|---|---|---|
| support.claude.com | 3 | |
| platform.claude.com | 2 | |
| docs.github.com | 1 | |
| learn.microsoft.com | 1 |