7 vs 15 benchmarks won
Anthropic Claude Opus 4.7 | OpenAI GPT-5.5 | |
|---|---|---|
| Overview | ||
| Company | Anthropic | OpenAI |
| Release date | Apr 16 2026 | Apr 23 2026 |
| Model type | — | — |
| Open source | No | No |
| Specifications | ||
Parameters | — | — |
Context window | — | — |
| Benchmarks | ||
Agentic coding SWE-Bench ProCan the AI fix real bugs in real software? It's handed actual problems from open-source projects and has to write code that genuinely solves them. Higher is better. | 64.3%Best | 58.6% |
Coding SWE-Bench VerifiedReal coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better. | 87.6% | — |
Multilingual coding SWE-Bench MultilingualLike SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better. | 80.5%Best | 77.8% |
Agentic coding CursorBench v3.1Cursor's own test of harder, real-world coding tasks inside a code editor. Higher is better. | 61.6%Best | 59.2% |
Agentic terminal coding Terminal-Bench 2.1Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Higher is better. | 66.1% | 78.2%Best |
Agentic terminal coding Terminal-Bench 2.0Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? (Version 2.0 of the test.) Higher is better. | 69.4% | 82.7%Best |
Software engineering Expert-SWE (Internal)OpenAI's private set of expert-level software-engineering problems. Higher is better. | — | 73.1% |
Multi-step tool use MCP AtlasCan the AI chain together many tools and steps to complete one bigger task, rather than doing just a single thing? Higher is better. | 79.1%Best | 75.3% |
General tool use ToolathlonTests how well the AI uses everyday real-world tools and apps to get things done. Higher is better. | — | 55.6% |
Web browsing BrowseCompCan the AI browse the web and track down hard-to-find answers? Higher is better. | 79.3% | 84.4%Best |
Cybersecurity CyberGymTests the AI on cybersecurity challenges — finding and exploiting software weaknesses inside a safe sandbox. Higher is better. | 73.1% | 81.8%Best |
Multidisciplinary reasoning Humanity's Last Exam · no toolsHumanity's Last Exam — extremely hard expert questions across many subjects, written so you can't just look up the answer. “No tools” means the AI answers on its own. Higher is better. | 46.9%Best | 41.4% |
Multidisciplinary reasoning Humanity's Last Exam · with toolsHumanity's Last Exam — extremely hard expert questions across many subjects. “With tools” means the AI is allowed to search the web or run code while answering. Higher is better. | 54.7%Best | 52.2% |
Abstract reasoning ARC-AGI-2Puzzle-style tests of abstract reasoning and pattern-finding — the kind of thing people find easy but AIs often struggle with. Higher is better. | 75.8% | 84.6%Best |
Advanced math FrontierMath · Tier 1–3Very hard, research-level math problems. Tiers 1–3 are the (still extremely difficult) lower tiers. Higher is better. | 43.8% | 51.7%Best |
Advanced math FrontierMath · Tier 4Very hard, research-level math problems. Tier 4 is the hardest — close to what professional research mathematicians tackle. Higher is better. | 22.9% | 35.4%Best |
Science GPQA DiamondGraduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 94.2%Best | 93.6% |
Agentic computer use OSWorld-VerifiedCan the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Higher is better. | 78% | 78.7%Best |
Agentic financial analysis Finance Agent v2Tests the AI on real financial-analysis work, like digging through reports and making sound decisions. Higher is better. | 51.5% | 51.8%Best |
Knowledge work GDPval-AAMeasures how well the AI does economically valuable knowledge work, judged against human experts. Shown as a rating (like a chess Elo) — higher is better. | 1753 | 1769Best |
Knowledge work GDPval (win/tie rate)How often the AI's work matches or beats a human expert's on real knowledge-work tasks. Higher is better. | 80.3% | 84.9%Best |
Chart reasoning CharXiv ReasoningCan the AI read and reason about complex charts and figures, not just text? Higher is better. | 82.1% | 84.1%Best |
Multimodal reasoning MMMU-ProA tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better. | 75.2% | 81.2%Best |
Spatial reasoning Blueprint-Bench 2Can the AI reason about space and layout — for example, understanding a floor plan or blueprint? Higher is better. | 24.5% | 36.2%Best |
Long context MRCR v2 (8-needle) · 128k averageTests whether the AI can find specific details buried inside a very long document (around 128k tokens — roughly a long book). Higher is better. | 59.3% | 94.8%Best |
| Timeline | ||
| Release gap | Claude Opus 4.7 shipped 7 days before GPT-5.5 | |
GPT-5.5 leads Claude Opus 4.7 on 15 of the 22 benchmarks they both report. Claude Opus 4.7 shipped 7 days before GPT-5.5, so benchmark comparisons should account for the intervening progress.
Published specifications for these two models are limited — see each model page for the latest details.
On SWE-Bench Pro, Claude Opus 4.7 leads at 64.3% vs GPT-5.5 at 58.6%. On SWE-Bench Multilingual, Claude Opus 4.7 leads at 80.5% vs GPT-5.5 at 77.8%. On CursorBench v3.1, Claude Opus 4.7 leads at 61.6% vs GPT-5.5 at 59.2%. On Terminal-Bench 2.1, GPT-5.5 leads at 78.2% vs Claude Opus 4.7 at 66.1%. On Terminal-Bench 2.0, GPT-5.5 leads at 82.7% vs Claude Opus 4.7 at 69.4%. On MCP Atlas, Claude Opus 4.7 leads at 79.1% vs GPT-5.5 at 75.3%. On BrowseComp, GPT-5.5 leads at 84.4% vs Claude Opus 4.7 at 79.3%. On CyberGym, GPT-5.5 leads at 81.8% vs Claude Opus 4.7 at 73.1%. On Humanity's Last Exam · no tools, Claude Opus 4.7 leads at 46.9% vs GPT-5.5 at 41.4%. On Humanity's Last Exam · with tools, Claude Opus 4.7 leads at 54.7% vs GPT-5.5 at 52.2%. On ARC-AGI-2, GPT-5.5 leads at 84.6% vs Claude Opus 4.7 at 75.8%. On FrontierMath · Tier 1–3, GPT-5.5 leads at 51.7% vs Claude Opus 4.7 at 43.8%. On FrontierMath · Tier 4, GPT-5.5 leads at 35.4% vs Claude Opus 4.7 at 22.9%. On GPQA Diamond, Claude Opus 4.7 leads at 94.2% vs GPT-5.5 at 93.6%. On OSWorld-Verified, GPT-5.5 leads at 78.7% vs Claude Opus 4.7 at 78%. On Finance Agent v2, GPT-5.5 leads at 51.8% vs Claude Opus 4.7 at 51.5%. On GDPval-AA, GPT-5.5 leads at 1769 vs Claude Opus 4.7 at 1753. On GDPval (win/tie rate), GPT-5.5 leads at 84.9% vs Claude Opus 4.7 at 80.3%. On CharXiv Reasoning, GPT-5.5 leads at 84.1% vs Claude Opus 4.7 at 82.1%. On MMMU-Pro, GPT-5.5 leads at 81.2% vs Claude Opus 4.7 at 75.2%. On Blueprint-Bench 2, GPT-5.5 leads at 36.2% vs Claude Opus 4.7 at 24.5%. On MRCR v2 (8-needle) · 128k average, GPT-5.5 leads at 94.8% vs Claude Opus 4.7 at 59.3%.