Across 31 comparable benchmarks, GPT-5.5 has better reported results on 31 and worse on 0. 2 further shared results are listed below, each with the reason it is not counted.
Moderate evidence31 of 33 shared results countedHow we compare
Price basis: Listed rates vs standard rates. The price section below shows the applicable periods.
| Fact | GPT-5.5 | Claude Sonnet 4.5 |
|---|---|---|
| Context window | 1M tokens | 1M tokens |
| Input / 1M | $5.00 | $3.00 |
| Output / 1M | $30.00 | $15.00 |
| Blended (4:1) | $6.76 | $3.46 |
| Status | Available | Deprecated |
| Released | April 23, 2026 | September 30, 2025 |
Blended is 4:1 input to output with 90% cache hits at listed rates — set your own mix.
Mean of 22 shared benchmarks grouped by capability on a 0–100 scale. How averages are built
GPT-5.5 ahead on 7 of 7 counted
GPT-5.5 ahead on 2 of 2 counted
GPT-5.5 ahead on 4 of 4 counted
GPT-5.5 ahead on 5 of 5 counted
GPT-5.5 ahead on 2 of 2 counted
GPT-5.5 ahead on 1 of 1 counted
GPT-5.5 ahead on 1 of 1 counted
GPT-5.5 ahead on 2 of 2 counted
GPT-5.5 ahead on 3 of 3 counted
No shared comparable results for Cybersecurity.
Published scores, with matching modes where available. Gaps show GPT-5.5 relative to Claude Sonnet 4.5. Reading the results
| Benchmark | GPT-5.5 | Claude Sonnet 4.5 | Gap |
|---|---|---|---|
| APEX-SWERepository Engineering · Pass@1 | 37.0Extra-High · tools | 31.0≠Thinking High · tools | +6.0 pts#38 vs #50 of 61 |
| GSOMaintenance & Optimization · OPT@1 (%) | 40.2Extra-High · tools | 14.7≠Normal · tools | +25.5 pts#8 vs #15 of 24 |
| SWE-Bench Pro - PublicRepository Engineering · Accuracy | 58.6Thinking High · tools | 43.6≠Normal · tools | +15.0 pts#22 vs #56 of 66 |
| Terminal Bench 2.0Agentic Development · Accuracy | 82.7Thinking High · tools | 42.8≠Thinking · tools | +39.9 pts#1 vs #41 of 46 |
| Terminal Bench HardAgentic Development · Accuracy | 60.6Extra-High · tools | 35.6≠Thinking · tools | +25.0 pts#4 vs #47 of 157 |
| Vibe Code Bench v1.1Code Generation & Editing · Pass rate (%) | 69.8Extra-High · tools | 22.6≠Thinking · tools | +47.2 pts#50 vs #83 of 119 |
| WeirdML v2ML Engineering · Average accuracy across 17 tasks (%) | 84.9Extra-High · tools | 47.7≠Thinking · tools · 16K | +37.2 pts#4 vs #27 of 39 |
| ARC-AGI-1Abstract Generalization · Score (%) | 95.0Extra-High · no tools | 63.7≠Thinking · no tools | +31.3 pts#20 vs #55 of 86 |
| ARC-AGI-2Abstract Generalization · Score (% solved); cost per task (USD)Not counted · Speed and cost telemetry is reported, not scored | 85.0Extra-High · no tools | 13.6≠Thinking · no tools | 6.25×#14 vs #56 of 79 |
| SimpleBenchCommonsense · Score (AVG@5) | 69.0Normal · no tools | 54.3Normal · no tools | 1.27×#27 vs #54 of 102 |
| Vending-Bench 2Logical Planning · Money Balance (USD)Not counted · Speed and cost telemetry is reported, not scored | 7524Unspecified Reasoning · tools | 3839Unspecified Reasoning · tools | 1.96×#22 vs #45 of 64 |
| FrontierMathMathematics · Accuracy | 51.7Thinking High · tools | 5.2≠Normal · no tools | +46.5 pts#2 vs #29 of 48 |
| FrontierMath - Tier 4Mathematics · Accuracy | 35.4Thinking High · tools | 4.2≠Thinking · no tools · 32K | +31.2 pts#3 vs #21 of 50 |
| FrontierMath Tier 4 v2Mathematics · Accuracy (verification_code) | 72.5Extra-High · tools | 2.4≠Thinking · tools · 32K | +70.1 pts#16 vs #66 of 73 |
| FrontierMath v2Mathematics · Accuracy (verification_code) | 85.3Extra-High · tools | 23.9≠Thinking · tools · 32K | +61.4 pts#16 vs #76 of 85 |
| CritPtScientific Reasoning · Score | 27.1Extra-High · no tools | 1.1≠Thinking · no tools | 24.6×#18 vs #103 of 134 |
| GPQA DiamondScientific Reasoning · Accuracy | 94.0Extra-High · no tools | 73.7≠Normal · no tools | +20.3 pts#19 vs #124 of 200 |
| SciCodeScientific Computing · Score | 56.1Thinking High · no tools | 45.7≠Thinking · no tools | 1.23×#28 vs #61 of 98 |
| MedCodeClinical Workflows · Accuracy (%) | 49.1Extra-High · tools | 44.1≠Thinking · tools | +5.0 pts#26 vs #40 of 97 |
| MedScribeClinical Workflows · Rubric pass rate (%) | 86.9Extra-High · tools | 84.5≠Normal · tools | +2.4 pts#20 vs #33 of 99 |
| MCP-AtlasTool Orchestration · Pass rate / claim coverage | 75.3Extra-High · tools | 59.5≠Thinking · tools | +15.8 pts#30 vs #43 of 51 |
| OSWorld-VerifiedDesktop Workflows · Accuracy | 78.7Thinking High · tools | 62.9≠Unspecified Reasoning · tools | +15.8 pts#13 vs #27 of 36 |
| BrowseCompFact Finding · Accuracy | 84.4Thinking High · tools + web | 24.1≠Thinking · tools | +60.3 pts#14 vs #55 of 57 |
| EBR-benchMemory & Persistence · Score (% of 21 objectives, best of the two scored playthroughs) | 34.3Extra-High · tools | 2.4≠Normal · tools | +31.9 pts#14 vs #29 of 30 |
| GDP.pdfDocuments & Charts · Task pass rate (%) | 22.6Thinking High · no tools | 5.2≠Thinking · no tools | +17.4 pts#26 vs #71 of 93 |
| MMMU-ProVisual Understanding · Accuracy | 81.2Thinking Medium · no tools | 68.9≠Reported best | +12.3 pts#26 vs #87 of 134 |
| Creative WritingWriting | 1845Normal · no tools | 1677Normal · no tools | 1.10×#17 vs #36 of 121 |
| SAGEService Workflows · Accuracy (%) | 51.5Extra-High · tools | 36.1≠Thinking · tools | +15.5 pts#13 vs #64 of 82 |
| τ³-BankingService Workflows · Score | 44.6Extra-High · tools | 24.5≠Thinking · tools | 1.82×#15 vs #49 of 120 |
| HLEKnowledge Exams · Accuracy | 52.2Thinking High · tools | 33.6≠Thinking · tools | +18.6 pts#40 vs #84 of 151 |
| ECICapability Indices · ECI score (capability index, higher is better) | 159Reported best | 147Reported best | 1.08×#10 vs #63 of 168 |
| LiveBenchCross-capability Suites · Accuracy | 80.2Extra-High · no tools | 68.2≠Thinking · no tools · 64K | +12.0 pts#10 vs #65 of 106 |
| Text Arena (Coding)Preference Arenas · Arena Score | 1505Extra-High · no tools | 1391≠Thinking · no tools · 32K | 1.08×#17 vs #26 of 28 |
Repository Engineering · Pass@1
Maintenance & Optimization · OPT@1 (%)
Repository Engineering · Accuracy
Agentic Development · Accuracy
Agentic Development · Accuracy
Code Generation & Editing · Pass rate (%)
ML Engineering · Average accuracy across 17 tasks (%)
Abstract Generalization · Score (%)
Abstract Generalization · Score (% solved); cost per task (USD)
Not counted · Speed and cost telemetry is reported, not scored
Commonsense · Score (AVG@5)
Logical Planning · Money Balance (USD)
Not counted · Speed and cost telemetry is reported, not scored
Mathematics · Accuracy
Mathematics · Accuracy
Mathematics · Accuracy (verification_code)
Mathematics · Accuracy (verification_code)
Scientific Reasoning · Score
Scientific Reasoning · Accuracy
Scientific Computing · Score
Clinical Workflows · Accuracy (%)
Clinical Workflows · Rubric pass rate (%)
Tool Orchestration · Pass rate / claim coverage
Desktop Workflows · Accuracy
Fact Finding · Accuracy
Memory & Persistence · Score (% of 21 objectives, best of the two scored playthroughs)
Writing
Service Workflows · Accuracy (%)
Service Workflows · Score
Knowledge Exams · Accuracy
Capability Indices · ECI score (capability index, higher is better)
Cross-capability Suites · Accuracy
Preference Arenas · Arena Score
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 · 5% steps
Blended price = effective input rate at 90% cache hits × 80% + output rate × 20%
GPT-5.5 costs96%morethan Claude Sonnet 4.5 per million tokens.Using listed rates vs standard rates at this token mix.
GPT-5.5: Standard pricing. context_length <= 272000 OpenAI official pricing (developers.openai.com/api/docs/pricing) verified 2026-10-21; standard tier; <=272K input tokens context_length > 272000 OpenAI official pricing (developers.openai.com/api/docs/pricing) verified 2026-10-21; standard tier; >272K input tokens; start date not published, verification date used
Claude Sonnet 4.5: Standard pricing. Anthropic official pricing (platform.claude.com/docs/en/about-claude/pricing) verified 2026-11-13; Deprecated 2026-09-30; tentative retirement 2026-11-30 (replacement: claude-sonnet-5-5)
| Price / 1M tokens | GPT-5.5 | Claude Sonnet 4.5 | Ratio |
|---|---|---|---|
| Input | $5.00 | $3.00 | 1.67× |
| Output | $30.00 | $15.00 | 2.00× |
| Cached input | $0.50 | $0.30 | 1.67× |
| Cache write | — | — | — |
| Blended | $6.76 | $3.46 | 1.96× |
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.
GPT-5.5 relative to Claude Sonnet 4.5
No matching recorded settings for a cost summary.
Summary: 0 of 2 shared benchmarks have matching recorded settings and sources. The average is a geometric mean of 0 positive-cost ratios, one equal weight per benchmark. This benchmark set can include related tests and indexes; it is not a workload cost estimate.
Settings or sources differ or are not recorded; interpret the comparison with care.
GPT-5.5
Claude Sonnet 4.5
Settings or sources differ or are not recorded; interpret the comparison with care.
GPT-5.5
Claude Sonnet 4.5
| Benchmark | GPT-5.5 | Claude Sonnet 4.5 |
|---|---|---|
| AA-LCR | Not reported | |
| Agents' Last Exam v1.0 | Not reported | |
| AIME 2026 | Not reported | |
| AutomationBench | $0.20 / task Average · Pass Rate: 1.1 Reported best (effort unspecified) · zapier-private-heldout $0.94 / task Average · Pass Rate: 12.8 Thinking High (With Tools) · zapier-private-heldout 2 more settings$0.75 / task Average · Pass Rate: 9.0 Thinking Medium (With Tools) · zapier-private-heldout | Not reported |
| DeepSWE | 1 more settings | Not reported |
| FrontierCode 1.1 | $2.6054 / task Average · Weighted rubric score: 54.0 Thinking High (With Tools) · frontiercode-1-1-extended $3.1187 / task Average · Weighted rubric score: 40.6 Thinking High (With Tools) · frontiercode-1-1-main $1.434 / task Average · Weighted rubric score: 45.5 Thinking Low (With Tools) · frontiercode-1-1-extended 5 more settings$1.6478 / task Average · Weighted rubric score: 30.6 Thinking Low (With Tools) · frontiercode-1-1-main $1.9847 / task Average · Weighted rubric score: 51.4 Thinking Medium (With Tools) · frontiercode-1-1-extended $2.3513 / task Average · Weighted rubric score: 36.6 Thinking Medium (With Tools) · frontiercode-1-1-main $4.0348 / task Average · Weighted rubric score: 43.0 Extra-High (With Tools) · frontiercode-1-1-main $3.3393 / task Average · Weighted rubric score: 56.7 Extra-High (With Tools) · frontiercode-1-1-extended | Not reported |
| FrontierCode 2.0 | $3.8021 / task Average · Weighted rubric score: 19.1 Thinking High (With Tools + Internet) $2.9515 / task Average · Weighted rubric score: 14.2 Thinking Medium (With Tools + Internet) 1 more settings | Not reported |
| LiveBench 2.0 | $2.2036 / task Average · Overall score (equal-weight mean of 7 suite scores, 0-100): 68.7 Extra-High (With Tools) | Not reported |
| Attribute | GPT-5.5 | Claude Sonnet 4.5 |
|---|---|---|
| Availability | Available | Deprecated |
| Provider | OpenAI | Anthropic |
| Model type | Reasoning model | Chat model |
| First release | April 23, 2026 | September 30, 2025 |
| Context window | 1M tokens | 1M tokens |
| Max output | 128K tokens | 64K tokens |
| Architecture | Undisclosed | Undisclosed |
| Documented capabilities | Reasoning, Multilingual, Coding, Tool use, Function calling, Computer use, Web search, Long context | Reasoning, Multilingual |
| 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 or Claude Sonnet 4.5? 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.