LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Claude Opus 4.8 vs Claude Opus 5

Updated October 21, 2026. 118 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Claude Opus 4.8 Replace Claude Opus 5
Claude Opus 4.8Anthropic
Coding
66.0
Overall
41.8
Output $/M
$25.00
4
Claude Opus 5Anthropic
Coding
81.8
Overall
50.8
Output $/M
$25.00
114
vs boards won of 118 shared

Decision reading

Pick Claude Opus 5 for coding: it scores 81.8 on the Vibe Coding Index against 66.0 for Claude Opus 4.8.

Pick Claude Opus 4.8 when what matters is LiveBench browser or Chess puzzles, where it leads.

Model AClaude Opus 4.8Anthropic66.0/1002 independent benchmarks$2.17 per solved taskModel BClaude Opus 5Anthropic81.8/1002 independent benchmarks$1.33 per solved task
Shared results
118
Claude Opus 4.8 only
8
Claude Opus 5 only
12
Like-for-like categories
8 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Claude Opus 4.8 66.0 from 2 benchmarks, Claude Opus 5 81.8 from 2 benchmarks.
Widest category
Math, 11.5 points apart on average over 6 shared benchmarks.
Widest single gap
Vending-Bench 2: Claude Opus 4.8 $5,787 against Claude Opus 5 $11,182, 33.7 points apart on a 0 to 100 scale.
Output price
Both cost $25.00 per million output tokens.
Context window
Both take 1M tokens.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryClaude Opus 4.8BarsClaude Opus 5Ties
Coding018
Intelligence02
Agentic work121
Reasoning116
Math06
Knowledge110
Web design09
General132
All shared boards4114
CodingLike-for-likeFour of these boards feed the Vibe Coding Index018
Coding benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
LiveBench coding77.3%90.8%Claude Opus 5
LMArena WebDev15561689Claude Opus 5
AA Coding Index74.378.0Claude Opus 5
LiveCodeBench (Vals AI)87.8%89.0%Claude Opus 5
SWE-bench Verified (Vals AI)88.6%97.0%Claude Opus 5
Terminal-Bench 4 (Vals AI)23.2%53.5%Claude Opus 5
Vibe Code Bench (Vals AI)82.7%88.4%Claude Opus 5
Code migration (Vals AI)47.3%57.5%Claude Opus 5
ProgramBench (Vals AI)1.0%3.0%Claude Opus 5
IOI (Vals AI)Not measured84.3%Only Claude Opus 5 measured
Terminal-Bench 2.1 (Vals AI)71.9%84.6%Claude Opus 5
Vibe Code Bench 1-100 (Vals AI)Not measured28.5%Only Claude Opus 5 measured
SkillsBench (Vals AI)59.2%60.4%Claude Opus 5
CyberBench (Vals AI)Not measured65.4%Only Claude Opus 5 measured
SciCode54.4%56.4%Claude Opus 5
Terminal-Bench Hard (Artificial Analysis run)58.3%Not measuredOnly Claude Opus 4.8 measured
FrontierCode46.5%53.4%Claude Opus 5
FrontierSWENot measured52.0%Only Claude Opus 5 measured
GSO-Bench47.1%Not measuredOnly Claude Opus 4.8 measured
WeirdML82.9%91.8%Claude Opus 5
ALE-BenchNot measured2164.6Only Claude Opus 5 measured
LMArena Text: coding14901533Claude Opus 5
LMArena Text: software and IT14881527Claude Opus 5
LMArena WebDev: React15571711Claude Opus 5
LMArena WebDev: HTML15431680Claude Opus 5
LMArena WebDev: from an imageNot measured1662Only Claude Opus 5 measured
Neither model has a result on the other 12 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
AA Intelligence Index41.850.8Claude Opus 5
Epoch Capabilities Index159.0163.9Claude Opus 5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index121
Agentic work benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
AA Agentic Index41.956.5Claude Opus 5
Design Arena: agents building web apps12331266Claude Opus 5
Design Arena: agents building full-stack apps12491306Claude Opus 5
Design Arena: agents building mobile apps12221348Claude Opus 5
Design Arena: agents building Android apps12371258Claude Opus 5
Design Arena: agents building Godot games12311279Claude Opus 5
Design Arena: agents building games12381261Claude Opus 5
Design Arena: HTML slides1221Not measuredOnly Claude Opus 4.8 measured
Design Arena: PowerPoint slides12271265Claude Opus 5
LiveBench research81.8%88.9%Claude Opus 5
LiveBench browser96.0%84.0%Claude Opus 4.8
LiveBench long agentic48.3%62.1%Claude Opus 5
SREBench (Vals AI)Not measured12.2%Only Claude Opus 5 measured
Finance Agent (Vals AI)53.9%58.6%Claude Opus 5
Tax Agent Bench (Vals AI)30.0%45.5%Claude Opus 5
Legal Research Bench (Vals AI)43.8%55.3%Claude Opus 5
Excel Modeling Benchmark (Vals AI)69.4%73.6%Claude Opus 5
BioMysteryBench (Vals AI)Not measured79.3%Only Claude Opus 5 measured
Terminal-Bench-Science (Vals AI)4.3%27.1%Claude Opus 5
Vals Science IndexNot measured47.9%Only Claude Opus 5 measured
Time Horizon Index (Vals AI)11.8%18.8%Claude Opus 5
CUA-bench (Vals AI)Not measured9.0%Only Claude Opus 5 measured
τ²-Bench Telecom94.4%Not measuredOnly Claude Opus 4.8 measured
APEX-Agents48.9%65.8%Claude Opus 5
Remote Labor Index8.3%Not measuredOnly Claude Opus 4.8 measured
OSWorld 2.020.6%31.4%Claude Opus 5
BALROGNot measured63.4%Only Claude Opus 5 measured
DeepResearch Bench50.2%Not measuredOnly Claude Opus 4.8 measured
PostTrainBench33.8%35.0%Claude Opus 5
Vending-Bench 2$5,787$11,182Claude Opus 5
LMArena Search12041242Claude Opus 5
Neither model has a result on the other 8 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index116
Reasoning benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
GPQA Diamond92.4%93.4%Claude Opus 5
LiveBench reasoning89.2%91.2%Claude Opus 5
MysteryMechanism (Vals AI)Not measured37.4%Only Claude Opus 5 measured
MMMU Pro (Vals AI)86.6%89.9%Claude Opus 5
GPQA Diamond (Artificial Analysis run)92.0%93.7%Claude Opus 5
Humanity's Last Exam (Artificial Analysis run)48.7%54.9%Claude Opus 5
GPQA Diamond (Epoch AI run)91.0%93.9%Claude Opus 5
ARC-AGI92.5%97.5%Claude Opus 5
ARC-AGI-272.1%90.4%Claude Opus 5
SimpleBench64.8%80.6%Claude Opus 5
Chess puzzles52.0%43.0%Claude Opus 4.8
Mystery game puzzles36.0%59.0%Claude Opus 5
EBR-bench28.6%45.7%Claude Opus 5
LMArena Text: hard prompts14821527Claude Opus 5
LMArena Text: expert15031554Claude Opus 5
LMArena Text: hard prompts in English14871523Claude Opus 5
LMArena Vision: diagrams13151348Claude Opus 5
LMArena Vision: homework13371338Claude Opus 5
Neither model has a result on the other 3 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index06
Math benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
OTIS Mock AIME98.3%98.9%Claude Opus 5
FrontierMath, tiers 1 to 380.0%85.6%Claude Opus 5
FrontierMath, tier 456.1%73.2%Claude Opus 5
ProofBench69.0%99.0%Claude Opus 5
LMArena Text: math14861533Claude Opus 5
LMArena Text: mathematical work14971544Claude Opus 5
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index110
Knowledge benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
MMLU Pro89.6%91.6%Claude Opus 5
LegalBench (Vals AI)83.6%87.0%Claude Opus 5
CorpFin (Vals AI)66.7%73.2%Claude Opus 5
TaxEval (Vals AI)75.6%75.1%Claude Opus 4.8
MortgageTax (Vals AI)69.9%72.1%Claude Opus 5
MedCode (Vals AI)53.2%63.6%Claude Opus 5
SimpleQA Verified53.0%59.9%Claude Opus 5
LMArena Text: science14771531Claude Opus 5
LMArena Text: medicine and healthcare14641525Claude Opus 5
LMArena Text: legal and government14771530Claude Opus 5
LMArena Text: business and finance14581501Claude Opus 5
Neither model has a result on the other 6 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index09
Web design benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
Design Arena: websites12581309Claude Opus 5
Design Arena: UI components12561348Claude Opus 5
Design Arena: game dev12641339Claude Opus 5
Design Arena: data viz12401329Claude Opus 5
Design Arena: 3D12231335Claude Opus 5
Design Arena: all code categories12551324Claude Opus 5
Design Arena: SVG11941310Claude Opus 5
Design Arena: ASCII art12761352Claude Opus 5
LiveBench design78.8%83.0%Claude Opus 5
GeneralLike-for-likePublic boards, shown as published, kept out of the index132
General benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
LiveBench data analysis66.0%74.6%Claude Opus 5
Public Benefits Bench (Vals AI)68.1%76.9%Claude Opus 5
SAGE (Vals AI)54.8%49.4%Claude Opus 4.8
MedScribe (Vals AI)85.8%91.0%Claude Opus 5
Vals Multimodal Index70.9%73.9%Claude Opus 5
IFBench62.2%Not measuredOnly Claude Opus 4.8 measured
Long context reasoning (AA-LCR)77.7%82.0%Claude Opus 5
DTBench94.9%97.9%Claude Opus 5
LMCA57.5%64.5%Claude Opus 5
Surface Evolver Bench87.5%Not measuredOnly Claude Opus 4.8 measured
Furniture assembly42.5%60.8%Claude Opus 5
LMArena Text14611509Claude Opus 5
LMArena Text: instruction following14761518Claude Opus 5
LMArena Text: creative writing14541493Claude Opus 5
LMArena Vision12941320Claude Opus 5
LMArena Text: English14681508Claude Opus 5
LMArena Text: non-English14501502Claude Opus 5
LMArena Text: Chinese15061573Claude Opus 5
LMArena Text: German14721527Claude Opus 5
LMArena Text: Japanese14421520Claude Opus 5
LMArena Text: Korean14311517Claude Opus 5
LMArena Text: French14811519Claude Opus 5
LMArena Text: Spanish14651523Claude Opus 5
LMArena Text: Russian14741507Claude Opus 5
LMArena Text: Polish14591522Claude Opus 5
LMArena Text: multi-turn14761500Claude Opus 5
LMArena Text: longer queries14831517Claude Opus 5
LMArena Text: writing and language14611499Claude Opus 5
LMArena Text: entertainment and media14401481Claude Opus 5
LMArena Vision: English12931315Claude Opus 5
LMArena Vision: Chinese13461400Claude Opus 5
LMArena Vision: reading text13031329Claude Opus 5
LMArena Vision: humor12821293Claude Opus 5
LMArena Vision: creative writing13011313Claude Opus 5
LMArena Document14751516Claude Opus 5
Neither model has a result on the other 11 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Opus 4.8 and Claude Opus 5
BenchmarkClaude Opus 4.8Claude Opus 5Leader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Claude Opus 4.8$0.018Fits in one request
Claude Opus 5$0.018Fits in one request

Same cost

Repository review

50K input + 3K output tokens

Claude Opus 4.8$0.325Fits in one request
Claude Opus 5$0.325Fits in one request

Same cost

Agent loop

200K input + 10K output tokens

Claude Opus 4.8$1.25Fits in one request
Claude Opus 5$1.25Fits in one request

Same cost

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Claude Opus 4.8 and Claude Opus 5
SpecificationAnthropic Claude Opus 4.8Anthropic Claude Opus 5
MakerAnthropicAnthropic
WeightsNo public weights listedNo public weights listed
ReleasedMay 27, 2026July 24, 2026
Context window1M1M
Input $/Mtok$5.00$5.00
Output $/Mtok$25.00$25.00
Coding index66.081.8
Independent measurements behind it22
Cost per solved task$2.17$1.33
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Opus 4.8 better than Claude Opus 5 for coding?

Pick Claude Opus 5 for coding: it scores 81.8 on the Vibe Coding Index against 66.0 for Claude Opus 4.8. Pick Claude Opus 4.8 when what matters is LiveBench browser or Chess puzzles, where it leads.

Which is cheaper, Claude Opus 4.8 or Claude Opus 5?

Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; Claude Opus 5 costs $5.00 and $25.00. Per solved SWE-bench Verified task, Claude Opus 4.8 costs $2.17 and Claude Opus 5 $1.33.

Which has the bigger context window, Claude Opus 4.8 or Claude Opus 5?

Neither: both take 1M tokens.

How many benchmarks is the Claude Opus 4.8 vs Claude Opus 5 comparison based on?

118 benchmarks have a result for both models as of October 21, 2026. Claude Opus 4.8's coding index rests on 2 independent benchmarks and Claude Opus 5's on 2. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/claude-opus-4-8-vs-claude-opus-5" title="Claude Opus 4.8 vs Claude Opus 5, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/claude-opus-4-8-vs-claude-opus-5">LLMBenchmark.ai, Claude Opus 4.8 vs Claude Opus 5</a></p>