Release Timeline
OpenAI unveiled GPT‑5.5 on April 23, 2026, just a week after Anthropic introduced its latest offering, Claude Opus 4.7. Both models are now available only to paid subscribers of their respective platforms.
Leaderboard Performance
GPT‑5.5 has not yet appeared on every AI leaderboard, but it is expected to compete closely with Claude Opus 4.7. On the Arc Prize verified benchmarks, GPT‑5.5 outperforms Opus 4.7. However, the popular Arena leaderboard—based on user testing—places Claude Opus 4.7 in the overall top spot, even though it is currently ranked below its predecessor, Opus 4.6. Anthropic’s unreleased Claude Mythos is not yet ranked but the company claims it surpasses Opus 4.7.
On the Epoch Capabilities Index (ECI), GPT‑5.4 Pro holds the highest score to date, followed by Gemini 3.1 Pro and GPT‑5.4. GPT‑5.5 and Opus 4.7 are not yet listed on the ECI.
Benchmark Results
The following table shows self‑reported scores from OpenAI and Anthropic for key benchmark tests. Percentages represent correct responses or overall performance as reported by each company.
| Benchmark | GPT‑5.5 | Claude Opus 4.7 |
|---|---|---|
| SWE‑Bench Pro | 58.6% | 64.3% |
| Terminal‑Bench 2.0 | 82.7% | 69.4% |
| Humanity’s Last Exam | 40.6% | 31.2% |
| Humanity’s Last Exam (with tools) | 52.2% | 54.7% |
| BrowseComp | 84.4% | 79.3% |
| GPQA Diamond | 93.6% | 94.2% |
| ARC‑AGI‑1 (Verified) | 94.5% (High) | 92% (High) |
| ARC‑AGI‑2 (Verified) | 83.3% (High) | 68.3% (High) |
Anthropic reports Opus 4.7 scored 46.9% on Humanity’s Last Exam, while the verified result from Artificial Analysis is listed above.
Availability and Pricing
OpenAI markets GPT‑5.5 as its “smartest and most intuitive” model yet. Claude Opus 4.7 is Anthropic’s flagship model for Claude users; the company says the unreleased Claude Mythos Preview will be even more capable.
GPT‑5.5 can be accessed by OpenAI Plus, Pro, Business, and Enterprise subscribers in ChatGPT and Codex (not available to ChatGPT Go users). Pro, Business, and Enterprise plans also include GPT‑5.5 Pro, while all paid tiers gain access to GPT‑5.5 Thinking.
API pricing for GPT‑5.5 starts at $5 per million input tokens and $30 per million output tokens, with a 1 million‑token context window. Claude Opus 4.7 is available to Pro and Max customers; its API costs $5 per million input tokens and $25 per million output tokens.
Feature Comparison
OpenAI highlights improvements in agentic coding, computer use, knowledge work, and early scientific research for GPT‑5.5. Anthropic emphasizes advanced coding, visual intelligence, and document analysis for Opus 4.7.
Both models support research, coding, creative projects, and general professional tasks. They are also integrated into their respective coding platforms—Codex for OpenAI and Claude Code for Anthropic.
Key differences include image generation: GPT‑5.5 users can employ ChatGPT Images 2.0, whereas Claude Design offers only data visualizations, graphics, and slides. GPT‑5.5 also supports more extensive app and shopping integrations; Anthropic’s recent acquisition of OpenClaw gives it an edge in agentic capabilities.
Bottom Line
If the goal is broad professional use with a wide array of features, GPT‑5.5 tends to be the stronger choice. For specialized tasks like advanced and agentic coding, Claude Opus 4.7 currently leads.
