Skip to main content

Model comparison

Claude Mythos 5 vs GPT-5.2

Data verified

Head-to-head evidence from 6 shared benchmark results across 4 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

83.93/100
Margin
25.5pts
← winning
OpenAI
58.43/100
4 category wins1 category wins

Public leaderboard positions: Claude Mythos 5 #1 (Supported); GPT-5.2 #64 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Claude Mythos 5 and GPT-5.2 share 6 comparable benchmark results. 5 of 8 categories are comparable. 9 results are unique to Claude Mythos 5; 22 to GPT-5.2.

Updated July 23, 2026
Shared results
6
Claude Mythos 5 only
9
GPT-5.2 only
22
Comparable categories
5 / 8

Pick Claude Mythos 5 if you want the stronger benchmark profile. GPT-5.2 only becomes the better choice if knowledge is the priority or you want the cheaper token bill.

Confidence note. This is a partial-evidence comparison with 6 shared benchmark results across 4 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

Claude Mythos 5 is clearly ahead on the BenchAlign aggregate, 83.93 to 58.43. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

Claude Mythos 5's sharpest advantage is in mathematics, where it averages 97.6 against 35.2. The single biggest benchmark swing on the page is OSWorld-Verified, 85% to 47.3%. GPT-5.2 does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.

Claude Mythos 5 is also the more expensive model on tokens at $10.00 input / $50.00 output per 1M tokens, versus $1.75 input / $14.00 output per 1M tokens for GPT-5.2. That is roughly 3.6x on output cost alone. Claude Mythos 5 gives you the larger context window at 1M+, compared with 400K for GPT-5.2.

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 Mythos 5 and GPT-5.2
CategoryClaude Mythos 5ΔGPT-5.2
MathClaude Mythos 597.6Margin 62.4GPT-5.235.2
AgenticClaude Mythos 587.0Margin 31.3GPT-5.255.7
KnowledgeClaude Mythos 568.5Margin 23.9GPT-5.292.4
CodingClaude Mythos 589.7Margin 19.1GPT-5.270.6
MultimodalClaude Mythos 593.5Margin 13.1GPT-5.280.4
ReasoningClaude Mythos 5Not measuredMarginNo overlapGPT-5.252.9

Decisive benchmark drivers

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

More
A · Claude Mythos 5B · GPT-5.2
  1. OSWorld-Verified

    Agentic
    Source ↗
    A 85%B 47.3%
    Winner: Claude Mythos 5Δ 37.7
    OSWorld-Verified: Claude Mythos 5 scored 85%; GPT-5.2 scored 47.3%. Claude Mythos 5 wins this benchmark.
  2. SWE-bench Pro

    Coding
    Source ↗
    A 80.3%B 55.6%
    Winner: Claude Mythos 5Δ 24.7
    SWE-bench Pro: Claude Mythos 5 scored 80.3%; GPT-5.2 scored 55.6%. Claude Mythos 5 wins this benchmark.
  3. BrowseComp

    Agentic
    Source ↗
    A 88%B 65.8%
    Winner: Claude Mythos 5Δ 22.2
    BrowseComp: Claude Mythos 5 scored 88%; GPT-5.2 scored 65.8%. Claude Mythos 5 wins this benchmark.
  4. SWE-bench Verified

    Coding
    Source ↗
    A 95.5%B 80%
    Winner: Claude Mythos 5Δ 15.5
    SWE-bench Verified: Claude Mythos 5 scored 95.5%; GPT-5.2 scored 80%. Claude Mythos 5 wins this benchmark.
  5. CharXiv

    Multimodal
    Source ↗
    A 93.5%B 82.1%
    Winner: Claude Mythos 5Δ 11.4
    CharXiv: Claude Mythos 5 scored 93.5%; GPT-5.2 scored 82.1%. Claude Mythos 5 wins this benchmark.

Operational comparison

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

MetricClaude Mythos 5GPT-5.2Comparison
Input / output priceUSD per 1M tokensClaude Mythos 5$10 input / $50 outputGPT-5.2$1.75 input / $14 outputGPT-5.2 has the lower combined listed price.
Generation speedtokens per secondClaude Mythos 5Not availableGPT-5.273 tok/sA complete speed comparison is not available.
First-answer latencyseconds to first tokenClaude Mythos 5Not availableGPT-5.2130.34 sA complete latency comparison is not available.
Context windowmaximum listed tokensClaude Mythos 51M+GPT-5.2400KClaude Mythos 5 lists the larger context window.

Benchmark Deep Dive

AgenticClaude Mythos 5 wins
BenchmarkClaude Mythos 5GPT-5.2Result
Terminal-Bench 2.0Source 88%Not comparable
OSWorld-VerifiedSource 85%47.3%Claude Mythos 5 leads
BrowseCompSource 88%65.8%Claude Mythos 5 leads
ExploitGymSource 17.5%Not comparable
τ²-bench resultsSource 84.8%Not comparable
Gert LabsSource 46.54%Not comparable
JobBenchSource 34.3%Not comparable
CodingClaude Mythos 5 wins
BenchmarkClaude Mythos 5GPT-5.2Result
SWE-bench VerifiedSource 95.5%80%Claude Mythos 5 leads
SWE-bench ProSource 80.3%55.6%Claude Mythos 5 leads
Terminal-Bench 2.0Source 88.0%Not comparable
Vibe Code BenchSource 53.50%Not comparable
AA-SciCodeSource 52.1%Not comparable
Reasoning
BenchmarkClaude Mythos 5GPT-5.2Result
ARC-AGI-2Source 52.9%Not comparable
AA-LCRSource 72.7%Not comparable
CritPtSource 11.6%Not comparable
KnowledgeGPT-5.2 wins
BenchmarkClaude Mythos 5GPT-5.2Result
GPQASource 94.1%92.4%Claude Mythos 5 leads
HLESource 64.5%Not comparable
HLE w/o toolsSource 59%Not comparable
Artificial Analysis Intelligence IndexSource 42.2%Not comparable
AA-GPQA DiamondSource 90.3%Not comparable
AA-HLESource 35.4%Not comparable
AA-Omniscience IndexSource -1.0%Not comparable
AA-Omniscience AccuracySource 43.8%Not comparable
AA-Omniscience Hallucination RateSource 79.7%Not comparable
MathClaude Mythos 5 wins
BenchmarkClaude Mythos 5GPT-5.2Result
USAMO 2026Source 97.6%Not comparable
AA AIME 2025Source 99.0%Not comparable
FrontierMath v2 (Tiers 1-3)Source 40.700%Not comparable
FrontierMath v2 (Tier 4)Source 18.800%Not comparable
Multilingual
BenchmarkClaude Mythos 5GPT-5.2Result
SWE MultilingualSource 92.2%Not comparable
MultimodalClaude Mythos 5 wins
BenchmarkClaude Mythos 5GPT-5.2Result
SWE-bench MultimodalSource 54.9%Not comparable
CharXivSource 93.5%82.1%Claude Mythos 5 leads
CharXiv w/o toolsSource 88.9%Not comparable
MMMU-ProSource 79.5%Not comparable
MathVisionSource 83.0%Not comparable
V*Source 75.9%Not comparable
Design Arena WebsiteSource 1224Not comparable
Inst. Following
BenchmarkClaude Mythos 5GPT-5.2Result
AA-IFBenchSource 75.4%Not comparable
Frequently Asked Questions (6)

Which is better, Claude Mythos 5 or GPT-5.2?

Claude Mythos 5 is ahead on BenchLM's BenchAlign leaderboard, 83.93 to 58.43. The biggest single separator in this matchup is OSWorld-Verified, where the scores are 85% and 47.3%.

Which is better for knowledge tasks, Claude Mythos 5 or GPT-5.2?

GPT-5.2 has the edge for knowledge tasks in this comparison, averaging 92.4 versus 68.5. Inside this category, GPQA is the benchmark that creates the most daylight between them.

Which is better for coding, Claude Mythos 5 or GPT-5.2?

Claude Mythos 5 has the edge for coding in this comparison, averaging 89.7 versus 70.6. Inside this category, SWE-bench Pro is the benchmark that creates the most daylight between them.

Which is better for math, Claude Mythos 5 or GPT-5.2?

Claude Mythos 5 has the edge for math in this comparison, averaging 97.6 versus 35.2. GPT-5.2 stays close enough that the answer can still flip depending on your workload.

Which is better for agentic tasks, Claude Mythos 5 or GPT-5.2?

Claude Mythos 5 has the edge for agentic tasks in this comparison, averaging 87 versus 55.7. Inside this category, OSWorld-Verified is the benchmark that creates the most daylight between them.

Which is better for multimodal and grounded tasks, Claude Mythos 5 or GPT-5.2?

Claude Mythos 5 has the edge for multimodal and grounded tasks in this comparison, averaging 93.5 versus 80.4. Inside this category, CharXiv is the benchmark that creates the most daylight between them.

Related Comparisons

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.