Skip to main content
BenchLM
Advertise

Model comparison

Claude Sonnet 4.6 vs GPT-5.6 Terra

Data verified

Head-to-head evidence from 18 shared benchmark results across 7 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

65.07/100
Margin
7.5pts
winning →
72.57/100
1 category wins4 category wins

Public leaderboard positions: Claude Sonnet 4.6 #32 (Supported); GPT-5.6 Terra #11 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Claude Sonnet 4.6 and GPT-5.6 Terra share 18 comparable benchmark results. 5 of 8 categories are comparable. 15 results are unique to Claude Sonnet 4.6; 26 to GPT-5.6 Terra.

Updated July 23, 2026
Shared results
18
Claude Sonnet 4.6 only
15
GPT-5.6 Terra only
26
Comparable categories
5 / 8

Pick GPT-5.6 Terra if you want the stronger benchmark profile. Claude Sonnet 4.6 only becomes the better choice if coding is the priority or you would rather avoid the extra latency and token burn of a reasoning model.

Confidence note. This is a partial-evidence comparison with 18 shared benchmark results across 7 evidence categories; 5 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

GPT-5.6 Terra is clearly ahead on the BenchAlign aggregate, 72.57 to 65.07. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

GPT-5.6 Terra's sharpest advantage is in mathematics, where it averages 80.8 against 26.4. The single biggest benchmark swing on the page is FrontierMath v2 (Tier 4), 8.300% to 68.300%. Claude Sonnet 4.6 does hit back in coding, so the answer changes if that is the part of the workload you care about most.

GPT-5.6 Terra is the reasoning model in the pair, while Claude Sonnet 4.6 is not. That usually helps on harder chain-of-thought-heavy tests, but it can also mean more latency and more token spend in real use. GPT-5.6 Terra gives you the larger context window at 1M, compared with 200K for Claude Sonnet 4.6.

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 4.6 and GPT-5.6 Terra
CategoryClaude Sonnet 4.6ΔGPT-5.6 Terra
MathClaude Sonnet 4.626.4Margin→ 54.4GPT-5.6 Terra80.8
KnowledgeClaude Sonnet 4.666.0Margin→ 26.9GPT-5.6 Terra92.9
AgenticClaude Sonnet 4.665.2Margin→ 22.2GPT-5.6 Terra87.4
CodingClaude Sonnet 4.669.1Margin← 5.7GPT-5.6 Terra63.4
MultimodalClaude Sonnet 4.677.4Margin→ 3.3GPT-5.6 Terra80.7

Decisive benchmark drivers

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

More
A · Claude Sonnet 4.6B · GPT-5.6 Terra

  1. FrontierMath v2 (Tier 4)

    Math
    Source ↗
    A 8.300%B 68.300%
    Winner: GPT-5.6 TerraΔ 60
    FrontierMath v2 (Tier 4): Claude Sonnet 4.6 scored 8.300%; GPT-5.6 Terra scored 68.300%. GPT-5.6 Terra wins this benchmark.

  2. FrontierMath v2 (Tiers 1-3)

    Math
    Source ↗
    A 32.400%B 84.900%
    Winner: GPT-5.6 TerraΔ 52.5
    FrontierMath v2 (Tiers 1-3): Claude Sonnet 4.6 scored 32.400%; GPT-5.6 Terra scored 84.900%. GPT-5.6 Terra wins this benchmark.

  3. Terminal-Bench 2.0

    Agentic
    Source ↗
    A 59.1%B 87.4%
    Winner: GPT-5.6 TerraΔ 28.3
    Terminal-Bench 2.0: Claude Sonnet 4.6 scored 59.1%; GPT-5.6 Terra scored 87.4%. GPT-5.6 Terra wins this benchmark.

  4. GPQA

    Knowledge
    Source ↗
    A 89.9%B 92.9%
    Winner: GPT-5.6 TerraΔ 3
    GPQA: Claude Sonnet 4.6 scored 89.9%; GPT-5.6 Terra scored 92.9%. GPT-5.6 Terra wins this benchmark.

Operational comparison

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

MetricClaude Sonnet 4.6GPT-5.6 TerraComparison
Input / output priceUSD per 1M tokensClaude Sonnet 4.6$3 input / $15 outputGPT-5.6 Terra$2.5 input / $15 outputGPT-5.6 Terra has the lower combined listed price.
Generation speedtokens per secondClaude Sonnet 4.644 tok/sGPT-5.6 TerraNot availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenClaude Sonnet 4.61.48 sGPT-5.6 TerraNot availableA complete latency comparison is not available.
Context windowmaximum listed tokensClaude Sonnet 4.6200KGPT-5.6 Terra1MGPT-5.6 Terra lists the larger context window.

Benchmark Deep Dive

AgenticGPT-5.6 Terra wins
BenchmarkClaude Sonnet 4.6GPT-5.6 TerraResult
Terminal-Bench 2.0Source 59.1%87.4%GPT-5.6 Terra leads
OSWorld-VerifiedSource 72.1%—Not comparable
Claw-EvalSource 67.8%—Not comparable
CyberGymSource 65.2%81.8%GPT-5.6 Terra leads
τ²-bench resultsSource 79.5%86.3%GPT-5.6 Terra leads
Gert LabsSource 62.92%—Not comparable
OSWorld 2.0Source 8.3%50.2%GPT-5.6 Terra leads
JobBenchSource 36.9%—Not comparable
BrowseCompSource —87.5%Not comparable
ExploitGymSource —23.2%Not comparable
ToolathlonSource —53.1%Not comparable
AA Agentic IndexSource —47.4%Not comparable
GDPval-AASource —54.1%Not comparable
GDPval-AASource —1581Not comparable
AA Harvey LABSource —85.2%Not comparable
AA ITBenchSource —51.0%Not comparable
AA Tau3 BankingSource —31.8%Not comparable
AA AutomationBenchSource —45.6%Not comparable
terminalBenchHardSource —57.6%Not comparable
aaTerminalBench21Source —88%Not comparable
APEX-Agents-AASource —38.9%Not comparable
CodingClaude Sonnet 4.6 wins
BenchmarkClaude Sonnet 4.6GPT-5.6 TerraResult
SWE-bench VerifiedSource 79.6%—Not comparable
SWE-RebenchSource 60.7%—Not comparable
React Native EvalsSource 80.6%—Not comparable
Vibe Code BenchSource 51.48%—Not comparable
cursorBench31Source 48.8%—Not comparable
AA-SciCodeSource 46.9%53.9%GPT-5.6 Terra leads
FrontierCode 1.1 MainSource 24.3%—Not comparable
SWE-bench ProSource —63.4%Not comparable
Terminal-Bench 2.0Source —87.4%Not comparable
deepSweSource —69.6%Not comparable
FrontierCode 1.1 ExtendedSource —55.8%Not comparable
cursorBench32Source —64.9%Not comparable
AA Coding IndexSource —76.7%Not comparable
Reasoning
BenchmarkClaude Sonnet 4.6GPT-5.6 TerraResult
AA-LCRSource 57.7%74.0%GPT-5.6 Terra leads
CritPtSource 0.9%30.0%GPT-5.6 Terra leads
ARC-AGI-3Source —0.8%Not comparable
KnowledgeGPT-5.6 Terra wins
BenchmarkClaude Sonnet 4.6GPT-5.6 TerraResult
GPQASource 89.9%92.9%GPT-5.6 Terra leads
SuperGPQASource 95%—Not comparable
MMLU-ProSource 79.2%—Not comparable
HLESource 49%—Not comparable
Artificial Analysis Intelligence IndexSource 35.9%55.0%GPT-5.6 Terra leads
AA-GPQA DiamondSource 79.9%92.5%GPT-5.6 Terra leads
AA-HLESource 13.2%41.8%GPT-5.6 Terra leads
AA-Omniscience IndexSource -2.9%-0.2%GPT-5.6 Terra leads
AA-Omniscience AccuracySource 38.0%45.9%GPT-5.6 Terra leads
AA-Omniscience Hallucination RateSource 65.9%85.2%Claude Sonnet 4.6 leads
GPQA-DSource —92.9%Not comparable
HealthBench ProfessionalSource —57.7%Not comparable
HealthBench HardSource —32.7%Not comparable
MathGPT-5.6 Terra wins
BenchmarkClaude Sonnet 4.6GPT-5.6 TerraResult
FrontierMath v2 (Tiers 1-3)Source 32.400%84.900%GPT-5.6 Terra leads
FrontierMath v2 (Tier 4)Source 8.300%68.300%GPT-5.6 Terra leads
FrontierMath (legacy)Source —84.9%Not comparable
MultimodalGPT-5.6 Terra wins
BenchmarkClaude Sonnet 4.6GPT-5.6 TerraResult
CharXivSource 77.4%—Not comparable
AA-MMMU-ProSource 70.6%80.7%GPT-5.6 Terra leads
Design Arena WebsiteSource 1314—Not comparable
MMMU-ProSource —80.7%Not comparable
MMMU-Pro w/ PythonSource —82%Not comparable
Inst. Following
BenchmarkClaude Sonnet 4.6GPT-5.6 TerraResult
AA-IFBenchSource 41.2%71.2%GPT-5.6 Terra leads
Frequently Asked Questions (6)

Which is better, Claude Sonnet 4.6 or GPT-5.6 Terra?

GPT-5.6 Terra is ahead on BenchLM's BenchAlign leaderboard, 72.57 to 65.07. The biggest single separator in this matchup is FrontierMath v2 (Tier 4), where the scores are 8.300% and 68.300%.

Which is better for knowledge tasks, Claude Sonnet 4.6 or GPT-5.6 Terra?

GPT-5.6 Terra has the edge for knowledge tasks in this comparison, averaging 92.9 versus 66. Inside this category, AA-HLE is the benchmark that creates the most daylight between them.

Which is better for coding, Claude Sonnet 4.6 or GPT-5.6 Terra?

Claude Sonnet 4.6 has the edge for coding in this comparison, averaging 69.1 versus 63.4. Inside this category, AA-SciCode is the benchmark that creates the most daylight between them.

Which is better for math, Claude Sonnet 4.6 or GPT-5.6 Terra?

GPT-5.6 Terra has the edge for math in this comparison, averaging 80.8 versus 26.4. Inside this category, FrontierMath v2 (Tier 4) is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, Claude Sonnet 4.6 or GPT-5.6 Terra?

GPT-5.6 Terra has the edge for agentic tasks in this comparison, averaging 87.4 versus 65.2. Inside this category, OSWorld 2.0 is the benchmark that creates the most daylight between them.

Which is better for multimodal and grounded tasks, Claude Sonnet 4.6 or GPT-5.6 Terra?

GPT-5.6 Terra has the edge for multimodal and grounded tasks in this comparison, averaging 80.7 versus 77.4. Inside this category, AA-MMMU-Pro 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.