BlogLLM Benchmark Scores as Published, Side by Side
Claude Opus 5: Benchmark Scores as Published
Every result the catalog holds for Claude Opus 5: 109 tests in 17 areas, from 20 pages. 67 printed by Anthropic itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.
Developer AnthropicReleased 2026-07-24Context window 1MMax output 128KInput Text, imageOutput TextWeights ClosedData version 2026103001, as of 2026-10-30
Knowledge Work
Agentic Coding
ML Engineering
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| PostTrainBench v1.1 | 40.4% | MiniMax, MiniMax-M3.1-Flash model cardchecked 2026-10-25 | Dataset revision: v1.1 |
Science and Math
Long Context
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| LOCA-Bench (256k environment description) | 53.7% | MiniMax, MiniMax-M3.1-Flash model cardchecked 2026-10-25 | - |
Computer Use
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| OSWorld-Verified | 83.4% | Xiaomi, Xiaomi MiMo-V2.6-Pro-RL model cardchecked 2026-10-19 | - |
| OSWorld 2.0 | 70.6% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 1 more page | - |
| OSWorld 2.0 partial | 74% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 3 more pages | Effort: max |
| OSWorld 2.1 partial | 74% | Anthropic, Claude Opus 5.5checked 2026-10-14 | - |
| OSWorld 2.0 strict | 37.2% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 2 more pages | Effort: max |
| OSWorld 2.0 offline partial | 70.2% | OpenAI, GPT-6 Astrachecked 2026-10-14; printed on 1 more page | - |
Multimodal Understanding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| CharXiv Reasoning (no tools) | 83.7% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15; printed on 1 more page | Tools: none |
| GDP.pdf | 21.6% | MiniMax, MiniMax-M3.1-Flash model cardchecked 2026-10-25 | - |
| BabyVision (with tools) | 94.1% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15; printed on 2 more pages | Tools: enabled; exact tools in source |
| LVBench | 75.4% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | - |
| Chartography (no tools) | 29.8% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 1 more page | Effort: max; Tools: none |
| Chartography (with tools) | 83.4% | Anthropic, Claude Opus 5.5checked 2026-10-14; printed on 4 more pages | Tools: enabled; exact tools in source |
| ZeroBench (with tools, pass@5) | 52% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15; printed on 1 more page | Tools: enabled; exact tools in source |
| GDP.pdf (all-pass rate) | 37% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | - |
Cybersecurity
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| ExploitBench | 70% | OpenAI, GPT-6 Astrachecked 2026-10-14; printed on 1 more page | - |
| ExploitGym | 22.1% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15; printed on 2 more pages | - |
| SRE-Bench | 12.5% | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 56.3% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 6 more pages | Tools: none |
| Humanity's Last Exam (with tools) | 64.7% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 6 more pages | Tools: enabled; exact tools in source |
| ARC-AGI-2 | 90.42% | Anthropic, System Card: Claude Fable 5.1 & Claude Mythos 5.1checked 2026-10-25; printed on 1 more page | Effort: max |
| ARC-AGI-1 | 97.5% | Anthropic, System Card: Claude Fable 5.1 & Claude Mythos 5.1checked 2026-10-25; printed on 1 more page | Effort: max |
| Humanity's Last Exam (Verified) | 54.4% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | - |
| ARC-AGI-3 | 30.2% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 1 more page | - |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Agents' Last Exam (pass rate) | 28.6% | DeepSeek, DeepSeek-V4.1-Flash model cardchecked 2026-10-15; printed on 3 more pages | - |
| Toolathlon Verified | 80.6% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 3 more pages | Effort: max |
| Agents' Last Exam (score) | 55.5% | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
| BankerToolBench | 76.2% | MiniMax, MiniMax-M3.1-Flash model cardchecked 2026-10-25 | - |
| Toolathlon Verified Pass@3 | 87% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 2 more pages | Effort: max |
| Toolathlon Verified Pass^3 | 73.1% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 2 more pages | Effort: max |
| Toolathlon Verified average assistant turnslower is better | 23.5 | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 1 more page | Effort: max |
Web Search and Research
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp | 90.8% | Anthropic, Claude Opus 5checked 2026-10-15; printed on 1 more page | - |
Multilingual
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Global-MMLU | 92.5% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 1 more page | Effort: max |
| MILU | 92.1% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 1 more page | Effort: max |
Healthcare
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| HealthBench Professional (length-adjusted) | 59.8% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 2 more pages | Effort: max |
| HealthBench Professional | 59.8% | Anthropic, Claude Opus 5checked 2026-10-15 | - |
| HealthBench (unadjusted) | 67.1 | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 1 more page | - |
| HealthBench Professional (unadjusted) | 73.4% | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 1 more page | - |
| PhysicianBench | 57.6% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | Effort: max |
Biosecurity
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BioSecBench Function | 44.8% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: max |
| BioSecBench Refusal | 31.8% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: max |
| BioSecBench Surveillance | 43.6% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: max |
Safety and Alignment
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Gray Swan IPI (attack success rate after k attempts)attack success rate after k attempts, lower is better | 4.6% | Google, Gemini 4 Argon Safety and Security Reportchecked 2026-10-19; printed on 1 more page | - |
| Self-preference bias, Claude-identity system prompt (0-9 scale)lower is better | -0.03 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
| Self-preference bias, no system prompt (0-9 scale)lower is better | 0.05 | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | - |
| Internal computer use safety benchmarklower is better | 11.5% | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
Engineering and Design
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BenchCAD | 82.1% | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
| EEBench reward | 61.6% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: max |
| BenchCAD Vision2Code, no tools (voxel IoU) | 0.497 | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 1 more page | Tools: none |
| BenchCAD Vision2Code, with tools (voxel IoU) | 0.899 | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25; printed on 1 more page | Tools: enabled |
| CADGenBench reward | 36.6% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: max |
Preference and Composite Rankings
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Artificial Analysis Intelligence Index v4.1.1 | 63.1 | OpenAI, GPT-6 Astrachecked 2026-10-14 | - |
| Anthropic ECI (AECI) | 165.18 | Anthropic, Claude Opus 5.5 System Cardchecked 2026-10-25 | - |
| Surge AI human eval on code (1 to 5) | 4.22 | Mistral AI, Mistral Large 4 (preview)checked 2026-10-15 | - |
Price
API list prices in US dollars per one million tokens, as printed on the developer's pricing page on the day checked. Cost per task is measured by Vals AI.
| Price | Value | Printed by | Setup as printed |
|---|---|---|---|
| Input | $5 | Anthropic, Anthropic API pricingchecked 2026-10-15 | - |
| Cached input | $0.5 | Anthropic, Anthropic API pricingchecked 2026-10-15 | - |
| Cache write | $6.25 | Anthropic, Anthropic API pricingchecked 2026-10-15 | Cache: 5 minute cache |
| Cache write, 1 hour | $10 | Anthropic, Anthropic API pricingchecked 2026-10-15 | - |
| Output | $25 | Anthropic, Anthropic API pricingchecked 2026-10-15 | - |
| Input, batch | $2.5 | Anthropic, Anthropic API pricingchecked 2026-10-15 | - |
| Output, batch | $12.5 | Anthropic, Anthropic API pricingchecked 2026-10-15 | - |
| Input, fast | $10 | Anthropic, Anthropic API pricingchecked 2026-10-15 | - |
| Output, fast | $50 | Anthropic, Anthropic API pricingchecked 2026-10-15 | - |
| Cost per taskVals AI, USD, lower is better | $33.88 | Vals AI, Vibe Code Bench v1.1 leaderboardchecked 2026-10-27 | Harness: OpenHands |
More from Anthropic
- Claude Haiku 5.5
- Claude Sonnet 5.5
- Claude Opus 5.5
- Claude Fable 5.1
- Claude Mythos 5.1
- Claude Sonnet 5
- Claude Fable 5
- Claude Mythos 5
- Claude Opus 4.8
- Claude Opus 4.7
- Claude Mythos Preview
- Claude Sonnet 4.6
- Claude Opus 4.6
- Claude Opus 4.5
- Claude Haiku 4.5
- Claude Sonnet 4.5
- Claude Opus 4.1
- Claude Opus 4
- Claude Sonnet 4
- Claude 3.7 Sonnet
- Claude 3.5 Sonnet
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.