Skip to main content
BenchLM
Advertise

Model comparison

Claude Sonnet 5 vs GPT-5.5

Data verified

Head-to-head evidence from 30 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

65.32/100
Margin
8.2pts
winning →
GPT-5.5
OpenAI
73.51/100
3 category wins1 category wins

Public leaderboard positions: Claude Sonnet 5 #29 (Estimated); GPT-5.5 #9 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Claude Sonnet 5 and GPT-5.5 share 30 comparable benchmark results. 4 of 8 categories are comparable. 6 results are unique to Claude Sonnet 5; 27 to GPT-5.5.

Updated July 23, 2026
Shared results
30
Claude Sonnet 5 only
6
GPT-5.5 only
27
Comparable categories
4 / 8

Pick GPT-5.5 if you want the stronger benchmark profile. Claude Sonnet 5 only becomes the better choice if coding is the priority or you want the cheaper token bill.

Confidence note. This is a partial-evidence comparison with 30 shared benchmark results across 5 evidence categories; 4 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

GPT-5.5 is clearly ahead on the BenchAlign aggregate, 73.51 to 65.32. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

GPT-5.5's sharpest advantage is in knowledge, where it averages 57.8 against 57.4. The single biggest benchmark swing on the page is HLE, 57.4% to 52.2%. Claude Sonnet 5 does hit back in coding, so the answer changes if that is the part of the workload you care about most.

GPT-5.5 is also the more expensive model on tokens at $5.00 input / $30.00 output per 1M tokens, versus $2.00 input / $10.00 output per 1M tokens for Claude Sonnet 5. That is roughly 3.0x on output cost alone.

Category breakdown

Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.

Category scores and score margins for Claude Sonnet 5 and GPT-5.5
CategoryClaude Sonnet 5ΔGPT-5.5
CodingClaude Sonnet 576.7Margin← 18.1GPT-5.558.6
MultimodalClaude Sonnet 588.3Margin← 17.9GPT-5.570.4
KnowledgeClaude Sonnet 557.4Margin→ 0.4GPT-5.557.8
AgenticClaude Sonnet 581.9Margin← 0.3GPT-5.581.6
ReasoningClaude Sonnet 5Not measuredMarginNo overlapGPT-5.585.0
MathClaude Sonnet 5Not measuredMarginNo overlapGPT-5.547.6

Decisive benchmark drivers

The largest measured benchmark gaps in this matchup, with exact reported values.

More
A · Claude Sonnet 5B · GPT-5.5

  1. HLE

    Knowledge
    Source ↗
    A 57.4%B 52.2%
    Winner: Claude Sonnet 5Δ 5.2
    HLE: Claude Sonnet 5 scored 57.4%; GPT-5.5 scored 52.2%. Claude Sonnet 5 wins this benchmark.

  2. SWE-bench Pro

    Coding
    Source ↗
    A 63.2%B 58.6%
    Winner: Claude Sonnet 5Δ 4.6
    SWE-bench Pro: Claude Sonnet 5 scored 63.2%; GPT-5.5 scored 58.6%. Claude Sonnet 5 wins this benchmark.

  3. OSWorld-Verified

    Agentic
    Source ↗
    A 81.2%B 78.7%
    Winner: Claude Sonnet 5Δ 2.5
    OSWorld-Verified: Claude Sonnet 5 scored 81.2%; GPT-5.5 scored 78.7%. Claude Sonnet 5 wins this benchmark.

  4. Terminal-Bench 2.0

    Agentic
    Source ↗
    A 80.4%B 82%
    Winner: GPT-5.5Δ 1.6
    Terminal-Bench 2.0: Claude Sonnet 5 scored 80.4%; GPT-5.5 scored 82%. GPT-5.5 wins this benchmark.

  5. BrowseComp

    Agentic
    Source ↗
    A 84.7%B 84.4%
    Winner: Claude Sonnet 5Δ 0.3
    BrowseComp: Claude Sonnet 5 scored 84.7%; GPT-5.5 scored 84.4%. Claude Sonnet 5 wins this benchmark.

Operational comparison

Runtime and commercial metrics are compared only when both models have a complete sourced value.

MetricClaude Sonnet 5GPT-5.5Comparison
Input / output priceUSD per 1M tokensClaude Sonnet 5$2 input / $10 outputGPT-5.5$5 input / $30 outputClaude Sonnet 5 has the lower combined listed price.
Generation speedtokens per secondClaude Sonnet 5Not availableGPT-5.5Not availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenClaude Sonnet 5Not availableGPT-5.5Not availableA complete latency comparison is not available.
Context windowmaximum listed tokensClaude Sonnet 51MGPT-5.51MListed context windows are equal.

Benchmark Deep Dive

AgenticClaude Sonnet 5 wins
BenchmarkClaude Sonnet 5GPT-5.5Result
Terminal-Bench 2.0Source 80.4%82%GPT-5.5 leads
BrowseCompSource 84.7%84.4%Claude Sonnet 5 leads
HLE w/ toolsSource 57.4%—Not comparable
OSWorld-VerifiedSource 81.2%78.7%Claude Sonnet 5 leads
GDPval-AASource 16071490Claude Sonnet 5 leads
AA Agentic IndexSource 46.7%44.9%Claude Sonnet 5 leads
GDPval-AASource 55.4%49.5%Claude Sonnet 5 leads
AA BriefcaseSource 13881154Claude Sonnet 5 leads
AA AutomationBenchSource 39.2%42.1%GPT-5.5 leads
AA EnterpriseOps-GymSource 44.7%46.6%GPT-5.5 leads
AA Harvey LABSource 90.1%86.3%Claude Sonnet 5 leads
AA Tau3 BankingSource 28.2%31.3%GPT-5.5 leads
aaTerminalBench21Source 80.5%84.3%GPT-5.5 leads
CyberGymSource —81.8%Not comparable
MCP AtlasSource —75.3%Not comparable
ToolathlonSource —55.6%Not comparable
τ²-bench resultsSource —93.9%Not comparable
APEX-Agents-AASource —37.7%Not comparable
Gert LabsSource —72.93%Not comparable
ResearchClawBenchSource —17.0%Not comparable
OSWorld 2.0Source —13.0%Not comparable
JobBenchSource —42.7%Not comparable
ExploitGymSource —13.4%Not comparable
AA ITBenchSource —45.8%Not comparable
terminalBenchHardSource —60.6%Not comparable
CodingClaude Sonnet 5 wins
BenchmarkClaude Sonnet 5GPT-5.5Result
SWE-bench VerifiedSource 85.2%—Not comparable
SWE-bench ProSource 63.2%58.6%Claude Sonnet 5 leads
SWE MultilingualSource 78.3%—Not comparable
SWE MultimodalSource 28.1%—Not comparable
Terminal-Bench 2.0Source 80.4%82.0%GPT-5.5 leads
FrontierCode 1.1 MainSource 42.7%43.0%GPT-5.5 leads
cursorBench32Source 61.5%58.4%Claude Sonnet 5 leads
AA Coding IndexSource 71.5%74.9%GPT-5.5 leads
AA-SciCodeSource 53.6%56.1%GPT-5.5 leads
Vibe Code BenchSource —69.85%Not comparable
React Native EvalsSource —84.7%Not comparable
cursorBench31Source —59.2%Not comparable
Reasoning
BenchmarkClaude Sonnet 5GPT-5.5Result
AA-LCRSource 70.7%74.3%GPT-5.5 leads
CritPtSource 16.9%27.1%GPT-5.5 leads
MRCR v2 64K-128KSource —83.1%Not comparable
MRCR v2 128K-256KSource —87.5%Not comparable
ARC-AGI-2Source —85%Not comparable
KnowledgeGPT-5.5 wins
BenchmarkClaude Sonnet 5GPT-5.5Result
HLESource 57.4%52.2%Claude Sonnet 5 leads
HLE w/o toolsSource 43.2%41.4%Claude Sonnet 5 leads
Artificial Analysis Intelligence IndexSource 53.4%54.8%GPT-5.5 leads
AA-GPQA DiamondSource 91.1%93.5%GPT-5.5 leads
AA-HLESource 39.6%44.3%GPT-5.5 leads
AA-Omniscience IndexSource 15.3%20.1%GPT-5.5 leads
AA-Omniscience AccuracySource 38.3%56.9%GPT-5.5 leads
AA-Omniscience Hallucination RateSource 37.3%85.5%Claude Sonnet 5 leads
GPQASource —93.6%Not comparable
GPQA-DSource —93.6%Not comparable
Math
BenchmarkClaude Sonnet 5GPT-5.5Result
FrontierMath (legacy)Source —51.7%Not comparable
FrontierMath v2 (Tiers 1-3)Source —51.700%Not comparable
FrontierMath v2 (Tier 4)Source —35.400%Not comparable
MultimodalClaude Sonnet 5 wins
BenchmarkClaude Sonnet 5GPT-5.5Result
CharXivSource 88.3%—Not comparable
CharXiv w/o toolsSource 77%—Not comparable
AA-MMMU-ProSource 77.3%79.9%GPT-5.5 leads
Design Arena WebsiteSource 13141282Claude Sonnet 5 leads
MMMU-ProSource —81.2%Not comparable
MMMU-Pro w/ PythonSource —83.2%Not comparable
OfficeQA ProSource —54.1%Not comparable
Inst. Following
BenchmarkClaude Sonnet 5GPT-5.5Result
AA-IFBenchSource —75.9%Not comparable
Frequently Asked Questions (5)

Which is better, Claude Sonnet 5 or GPT-5.5?

GPT-5.5 is ahead on BenchLM's BenchAlign leaderboard, 73.51 to 65.32. The biggest single separator in this matchup is HLE, where the scores are 57.4% and 52.2%.

Which is better for knowledge tasks, Claude Sonnet 5 or GPT-5.5?

GPT-5.5 has the edge for knowledge tasks in this comparison, averaging 57.8 versus 57.4. Inside this category, AA-Omniscience Hallucination Rate is the benchmark that creates the most daylight between them.

Which is better for coding, Claude Sonnet 5 or GPT-5.5?

Claude Sonnet 5 has the edge for coding in this comparison, averaging 76.7 versus 58.6. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, Claude Sonnet 5 or GPT-5.5?

Claude Sonnet 5 has the edge for agentic tasks in this comparison, averaging 81.9 versus 81.6. Inside this category, AA Briefcase is the benchmark that creates the most daylight between them.

Which is better for multimodal and grounded tasks, Claude Sonnet 5 or GPT-5.5?

Claude Sonnet 5 has the edge for multimodal and grounded tasks in this comparison, averaging 88.3 versus 70.4. Inside this category, Design Arena Website is the benchmark that creates the most daylight between them.

Related Comparisons

Explore More

Anthropic Compare Pricing Methodology Find Your Best LLM Overall Rankings
Last updated: July 23, 2026

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.