Across 8 comparable benchmarks, Claude Opus 5 has better reported results on 8 and worse on 0.
Moderate evidence8 of 8 shared results countedHow we compare
Price basis: Listed rates vs standard rates. The price section below shows the applicable periods.
| Fact | GPT-5.5 Instant | Claude Opus 5 |
|---|---|---|
| Context window | β | 1M tokens |
| Input / 1M | β | $5.00 |
| Output / 1M | β | $25.00 |
| Blended (4:1) | β | $5.76 |
| Status | Available | Available |
| Released | May 5, 2026 | July 24, 2026 |
Blended is 4:1 input to output with 90% cache hits at listed rates β set your own mix.
Mean of 4 shared benchmarks grouped by capability on a 0β100 scale. How averages are built
Claude Opus 5 ahead on 1 of 1 counted
Claude Opus 5 ahead on 2 of 2 counted
Claude Opus 5 ahead on 2 of 2 counted
No fixed-scale average β see the rows in Benchmark results
Claude Opus 5 ahead on 1 of 1 counted
Claude Opus 5 ahead on 1 of 1 counted
No fixed-scale average β see the rows in Benchmark results
No shared comparable results for Abstract reasoning, Computer & tool use, Research & search, Long context, Cybersecurity.
Published scores, with matching modes where available. Gaps show GPT-5.5 Instant relative to Claude Opus 5. Reading the results
| Benchmark | GPT-5.5 Instant | Claude Opus 5 | Gap |
|---|---|---|---|
| Terminal-Bench 4.0Agentic Development Β· Resolution rate (%) | 12.6Thinking Β· tools | 53.9β Extra-High Β· tools | β41.3 pts#58 vs #16 of 90 |
| FrontierMath Tier 4 v2Mathematics Β· Accuracy (verification_code) | 2.4Reported best | 73.2β Max Β· tools | β70.7 pts#66 vs #15 of 73 |
| FrontierMath v2Mathematics Β· Accuracy (verification_code) | 26.3Reported best | 85.6β Max Β· tools | β59.3 pts#73 vs #15 of 85 |
| CritPtScientific Reasoning Β· Score | 0.0Thinking Β· no tools | 29.1β Max Β· no tools | 0.00Γ#134 vs #13 of 134 |
| SciCodeScientific Computing Β· Score | 52.5Thinking Β· no tools | 56.4β Max Β· no tools | 0.93Γ#41 vs #27 of 98 |
| GDP.pdfDocuments & Charts Β· Task pass rate (%) | 20.2Thinking Β· no tools | 21.6β Max Β· no tools | β1.4 pts#29 vs #28 of 96 |
| ΟΒ³-BankingService Workflows Β· Score | 11.8Thinking Β· tools | 48.7β Max Β· tools | 0.24Γ#88 vs #6 of 120 |
| ECICapability Indices Β· ECI score (capability index, higher is better) | 142Reported best | 163Reported best | 0.88Γ#89 vs #6 of 168 |
Agentic Development Β· Resolution rate (%)
Mathematics Β· Accuracy (verification_code)
Mathematics Β· Accuracy (verification_code)
Documents & Charts Β· Task pass rate (%)
Service Workflows Β· Score
Capability Indices Β· ECI score (capability index, higher is better)
A dash means no result is available.
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix Β· 5% steps
Blended price = effective input rate at 90% cache hits Γ 80% + output rate Γ 20%
A blended cost needs a published input and output price on both sides, and one of these two does not have them yet.
GPT-5.5 Instant: Listed text rates; pricing tier not confirmed.
Claude Opus 5: Standard pricing. Anthropic official pricing (platform.claude.com/docs/en/about-claude/pricing) verified 2026-10-21; full 1M context at standard price
| Price / 1M tokens | GPT-5.5 Instant | Claude Opus 5 | Ratio |
|---|---|---|---|
| Input | β | $5.00 | β |
| Output | β | $25.00 | β |
| Cached input | β | $0.50 | β |
| Cache write | β | β | β |
| Blended | β | $5.76 | β |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
Reported cost per task, paired with the score from the same run. Settings and sources may differ; costs do not affect the verdict.
| Benchmark | GPT-5.5 Instant | Claude Opus 5 |
|---|---|---|
| AA Intelligence Index v4.3 | Not reported | 2 more settings |
| AA Intelligence Index v5.0.1 | Not reported | |
| AA-Omniscience | Not reported | 2 more settings |
| Agent Arena | Not reported | $4.6032 / task Median Β· Net improvement index (percentage points vs. average model): 7.1 Max (With Tools + Internet) |
| Agent Arena Index | Not reported | $4.6109 / task Median Β· Agent Arena Index rating (index points): 1176 Max (With Tools + Internet) |
| Agents' Last Exam v1.0 | Not reported | 2 more settings |
| ARC-AGI-1 | Not reported | |
| ARC-AGI-2 | Not reported | $1.45 / task Average Β· ARC-AGI-2 Score: 88.3 Thinking High (With Tools) |
| AutomationBench | Not reported | $1.03 / task Average Β· Pass Rate: 20.6 Thinking High (With Tools) Β· zapier-private-heldout $0.75 / task Average Β· Pass Rate: 20.4 Thinking Low (With Tools) Β· zapier-private-heldout 2 more settings$0.89 / task Average Β· Pass Rate: 23.9 Thinking Medium (With Tools) Β· zapier-private-heldout $1.15 / task Average Β· Pass Rate: 25.3 Extra-High (With Tools) Β· zapier-private-heldout |
| CursorBench 4.0 | Not reported | $11.95 / task Average Β· Task score: 46.6 Max (With Tools) Β· cursorbench-4-0-cursor-agent $9.00 / task Average Β· Task score: 44.7 Thinking High (With Tools) Β· cursorbench-4-0-cursor-agent $4.87 / task Average Β· Task score: 40.7 Thinking Low (With Tools) Β· cursorbench-4-0-cursor-agent 2 more settings$6.94 / task Average Β· Task score: 43.3 Thinking Medium (With Tools) Β· cursorbench-4-0-cursor-agent $11.43 / task Average Β· Task score: 46.1 Extra-High (With Tools) Β· cursorbench-4-0-cursor-agent |
| DeepSWE | Not reported | 2 more settings |
| FrontierCode 1.1 | Not reported | $11.416 / task Average Β· Weighted rubric score: 48.0 Max (With Tools) Β· frontiercode-1-1-main $7.2397 / task Average Β· Weighted rubric score: 48.0 Thinking High (With Tools) Β· frontiercode-1-1-main $2.678 / task Average Β· Weighted rubric score: 42.0 Thinking Low (With Tools) Β· frontiercode-1-1-main 2 more settings$4.3127 / task Average Β· Weighted rubric score: 53.4 Thinking Medium (With Tools) Β· frontiercode-1-1-main $9.1423 / task Average Β· Weighted rubric score: 43.6 Extra-High (With Tools) Β· frontiercode-1-1-main |
| FrontierCode 2.0 | Not reported | $8.361 / task Average Β· Weighted rubric score: 26.7 Thinking High (With Tools + Internet) $3.276 / task Average Β· Weighted rubric score: 18.7 Thinking Low (With Tools + Internet) 2 more settings$5.217 / task Average Β· Weighted rubric score: 24.0 Thinking Medium (With Tools + Internet) |
| FrontierSWE v2 | Not reported | |
| HLE | Not reported | $0.0625 / task Average Β· Accuracy: 43.4 Thinking Low (No Tools) Β· Text only Β· text-only $0.187 / task Average Β· Accuracy: 51.3 Thinking Medium (No Tools) Β· Text only Β· text-only 2 more settings$0.3521 / task Average Β· Accuracy: 52.8 Thinking High (No Tools) Β· Text only Β· text-only $0.5171 / task Average Β· Accuracy: 54.4 Extra-High (No Tools) Β· Text only Β· text-only |
| LiveBench 2.0 | Not reported | $3.6538 / task Average Β· Overall score (equal-weight mean of 7 suite scores, 0-100): 82.1 Max (task-specific tools) |
| SWE-sweep | Not reported | |
| Tax Agent Bench | Not reported | $5.1366 / task Average Β· Accuracy (%, rubric Γ citation quality): 45.5 Unspecified Reasoning (With Tools) |
| Terminal-Bench 4.0 | Not reported | 3 more settings$18.60 / task Average Β· Resolution rate (%): 53.5 Unspecified Reasoning (With Tools) Β· vals-mini-swe |
| Terminal-Bench 4.1 | Not reported | 2 more settings |
| Vals Index v2.1 | Not reported | |
| Vibe Code Bench 1-100 | Not reported | |
| WeirdML v3 | Not reported | $25.2622 / task Average Β· Official score (0-1; higher is better): 0.19 Extra-High (With Tools) |
| Attribute | GPT-5.5 Instant | Claude Opus 5 |
|---|---|---|
| Availability | Available | Available |
| Provider | OpenAI | Anthropic |
| Model type | Chat model | Reasoning model |
| First release | May 5, 2026 | July 24, 2026 |
| Context window | β | 1M tokens |
| Max output | β | 128K tokens |
| Architecture | Undisclosed | Undisclosed |
| Documented capabilities | β | Reasoning, Multilingual, Coding, Tool use, Function calling, Computer use, Long context |
| Knowledge cutoff | β | 2026-05 |
| Input modalities | Text, Image | Text, Image |
| Output modalities | Text | Text |
| Weights | Closed / API only | Closed / API only |
| Commercial use | Proprietary | Proprietary |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used GPT-5.5 Instant or Claude Opus 5? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.