Skip to main content
BenchLM
Advertise

Model comparison

Claude Opus 4.6 vs GPT-5.3 Codex

Data verified

Head-to-head evidence from 21 shared benchmark results across 6 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

68.59/100
Margin
1.9pts
← winning
66.69/100
2 category wins0 category wins

Public leaderboard positions: Claude Opus 4.6 #16 (Supported); GPT-5.3 Codex #26 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Claude Opus 4.6 and GPT-5.3 Codex share 21 comparable benchmark results. 2 of 8 categories are comparable. 25 results are unique to Claude Opus 4.6; 0 to GPT-5.3 Codex.

Updated July 20, 2026
Shared results
21
Claude Opus 4.6 only
25
GPT-5.3 Codex only
0
Comparable categories
2 / 8

Pick Claude Opus 4.6 if you want the stronger benchmark profile. GPT-5.3 Codex only becomes the better choice if you want the cheaper token bill or you want the stronger reasoning-first profile.

Confidence note. This is a partial-evidence comparison with 21 shared benchmark results across 6 evidence categories; 2 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

Claude Opus 4.6 has the cleaner BenchAlign overall profile here, landing at 68.59 versus 66.69. It is a real lead, but still close enough that category-level strengths matter more than the headline number.

Claude Opus 4.6's sharpest advantage is in agentic, where it averages 73 against 71.4. The single biggest benchmark swing on the page is Terminal-Bench 2.0, 65.4% to 77.3%.

Claude Opus 4.6 is also the more expensive model on tokens at $5.00 input / $25.00 output per 1M tokens, versus $1.75 input / $14.00 output per 1M tokens for GPT-5.3 Codex. GPT-5.3 Codex is the reasoning model in the pair, while Claude Opus 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. Claude Opus 4.6 gives you the larger context window at 1M, compared with 400K for GPT-5.3 Codex.

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 Opus 4.6 and GPT-5.3 Codex
CategoryClaude Opus 4.6ΔGPT-5.3 Codex
AgenticClaude Opus 4.673.0Margin← 1.6GPT-5.3 Codex71.4
CodingClaude Opus 4.668.1Margin← 0.9GPT-5.3 Codex67.2
KnowledgeClaude Opus 4.669.1MarginNo overlapGPT-5.3 CodexNot measured
MathClaude Opus 4.636.3MarginNo overlapGPT-5.3 CodexNot measured
MultimodalClaude Opus 4.677.3MarginNo overlapGPT-5.3 CodexNot measured

Decisive benchmark drivers

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

More
A · Claude Opus 4.6B · GPT-5.3 Codex

  1. Terminal-Bench 2.0

    Agentic
    Source ↗
    A 65.4%B 77.3%
    Winner: GPT-5.3 CodexΔ 11.9
    Terminal-Bench 2.0: Claude Opus 4.6 scored 65.4%; GPT-5.3 Codex scored 77.3%. GPT-5.3 Codex wins this benchmark.

  2. OSWorld-Verified

    Agentic
    Source ↗
    A 72.7%B 64.7%
    Winner: Claude Opus 4.6Δ 8
    OSWorld-Verified: Claude Opus 4.6 scored 72.7%; GPT-5.3 Codex scored 64.7%. Claude Opus 4.6 wins this benchmark.

  3. SWE-Rebench

    Coding
    Source ↗
    A 65.3%B 58.2%
    Winner: Claude Opus 4.6Δ 7.1
    SWE-Rebench: Claude Opus 4.6 scored 65.3%; GPT-5.3 Codex scored 58.2%. Claude Opus 4.6 wins this benchmark.

  4. SWE-bench Verified

    Coding
    Source ↗
    A 80.8%B 85%
    Winner: GPT-5.3 CodexΔ 4.2
    SWE-bench Verified: Claude Opus 4.6 scored 80.8%; GPT-5.3 Codex scored 85%. GPT-5.3 Codex wins this benchmark.

  5. SWE-bench Pro

    Coding
    Source ↗
    A 53.4%B 56.8%
    Winner: GPT-5.3 CodexΔ 3.4
    SWE-bench Pro: Claude Opus 4.6 scored 53.4%; GPT-5.3 Codex scored 56.8%. GPT-5.3 Codex wins this benchmark.

Operational comparison

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

MetricClaude Opus 4.6GPT-5.3 CodexComparison
Input / output priceUSD per 1M tokensClaude Opus 4.6$5 input / $25 outputGPT-5.3 Codex$1.75 input / $14 outputGPT-5.3 Codex has the lower combined listed price.
Generation speedtokens per secondClaude Opus 4.640 tok/sGPT-5.3 Codex79 tok/sGPT-5.3 Codex has the higher measured throughput.
First-answer latencyseconds to first tokenClaude Opus 4.61.78 sGPT-5.3 Codex88.26 sClaude Opus 4.6 reaches the first token sooner.
Context windowmaximum listed tokensClaude Opus 4.61MGPT-5.3 Codex400KClaude Opus 4.6 lists the larger context window.

Benchmark Deep Dive

AgenticClaude Opus 4.6 wins
BenchmarkClaude Opus 4.6GPT-5.3 CodexResult
Terminal-Bench 2.0Source 65.4%77.3%GPT-5.3 Codex leads
BrowseCompSource 83.7%—Not comparable
OSWorld-VerifiedSource 72.7%64.7%Claude Opus 4.6 leads
τ²-bench resultsSource 84.8%86%GPT-5.3 Codex leads
Claw-EvalSource 70.4%—Not comparable
DeepSearchQASource 73.7%—Not comparable
CyberGymSource 66.6%—Not comparable
Gert LabsSource 61.85%57.47%Claude Opus 4.6 leads
ResearchClawBenchSource 19.9%—Not comparable
JobBenchSource 36.7%33.7%Claude Opus 4.6 leads
CodingClaude Opus 4.6 wins
BenchmarkClaude Opus 4.6GPT-5.3 CodexResult
SWE-bench VerifiedSource 80.8%85%GPT-5.3 Codex leads
SWE-bench Verified*Source 75.6%—Not comparable
LiveCodeBench ProSource 70.7%—Not comparable
SWE-bench ProSource 53.4%56.8%GPT-5.3 Codex leads
SWE-RebenchSource 65.3%58.2%Claude Opus 4.6 leads
React Native EvalsSource 84.1%—Not comparable
Vibe Code BenchSource 57.57%61.77%GPT-5.3 Codex leads
AA-SciCodeSource 45.7%53.2%GPT-5.3 Codex leads
FrontierCode 1.1 MainSource 26.9%—Not comparable
Reasoning
BenchmarkClaude Opus 4.6GPT-5.3 CodexResult
AA-LCRSource 58.3%74.0%GPT-5.3 Codex leads
CritPtSource 2.8%16.9%GPT-5.3 Codex leads
Knowledge
BenchmarkClaude Opus 4.6GPT-5.3 CodexResult
GPQASource 91.3%—Not comparable
GPQA-DSource 89.2%—Not comparable
SuperGPQASource 95%—Not comparable
MMLU-ProSource 82%—Not comparable
MMLU-Pro (Arcee)Source 89.1%—Not comparable
HLESource 53%—Not comparable
HLE w/o toolsSource 40%—Not comparable
HealthBench HardSource 14.8%—Not comparable
MedXpertQA (Text)Source 52.1%—Not comparable
Artificial Analysis Intelligence IndexSource 37.8%44.3%GPT-5.3 Codex leads
AA-GPQA DiamondSource 84.0%91.5%GPT-5.3 Codex leads
AA-HLESource 18.6%39.9%GPT-5.3 Codex leads
AA-Omniscience IndexSource 3.5%9.9%GPT-5.3 Codex leads
AA-Omniscience AccuracySource 45.2%51.8%GPT-5.3 Codex leads
AA-Omniscience Hallucination RateSource 76.0%86.9%Claude Opus 4.6 leads
Math
BenchmarkClaude Opus 4.6GPT-5.3 CodexResult
AIME25 (Arcee)Source 99.8%—Not comparable
FrontierMath v2 (Tiers 1-3)Source 40.700%—Not comparable
FrontierMath v2 (Tier 4)Source 22.900%—Not comparable
Multimodal
BenchmarkClaude Opus 4.6GPT-5.3 CodexResult
MMMU-ProSource 77.3%—Not comparable
ERQASource 51.6%—Not comparable
ScreenSpot ProSource 83.1%—Not comparable
MedXpertQA (MM)Source 64.8%—Not comparable
AA-MMMU-ProSource 72.5%78.5%GPT-5.3 Codex leads
Design Arena WebsiteSource 13281195Claude Opus 4.6 leads
Inst. Following
BenchmarkClaude Opus 4.6GPT-5.3 CodexResult
AA-IFBenchSource 44.6%75.4%GPT-5.3 Codex leads
Frequently Asked Questions (3)

Which is better, Claude Opus 4.6 or GPT-5.3 Codex?

Claude Opus 4.6 is ahead on BenchLM's BenchAlign leaderboard, 68.59 to 66.69. The biggest single separator in this matchup is Terminal-Bench 2.0, where the scores are 65.4% and 77.3%.

Which is better for coding, Claude Opus 4.6 or GPT-5.3 Codex?

Claude Opus 4.6 has the edge for coding in this comparison, averaging 68.1 versus 67.2. Inside this category, AA-SciCode is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, Claude Opus 4.6 or GPT-5.3 Codex?

Claude Opus 4.6 has the edge for agentic tasks in this comparison, averaging 73 versus 71.4. Inside this category, Terminal-Bench 2.0 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 20, 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.