Enter opens the highlighted result. Esc closes.

Claude Opus 4.8 vs Claude Sonnet 5

Claude Opus 4.8 leads Claude Sonnet 5 on coding, 65.0 to 57.7 on the Vibe Coding Index, and Claude Sonnet 5 costs 2.5x less per output token ($10.00 vs $25.00 per million). Updated October 22, 2026; 115 benchmarks have a result for both.

Replace Claude Opus 4.8 Replace Claude Sonnet 5
Claude Opus 4.8Anthropic
Coding
65.0
Overall
69.8
Output $/M
$25.00
93
Claude Sonnet 5Anthropic
Coding
57.7
Overall
63.2
Output $/M
$10.00
20
vs boards won of 115 shared, 2 tied

Claude Opus 4.8 or Claude Sonnet 5: which to pick

Pick Claude Opus 4.8 for coding: it scores 65.0 on the Vibe Coding Index against 57.7 for Claude Sonnet 5.

Pick Claude Sonnet 5 for output tokens 2.5x cheaper ($10.00 vs $25.00 per million) and a lower cost per solved task ($1.87 against $2.17).

Model AClaude Opus 4.8Anthropic65.0/1006 independent benchmarks$2.17 per solved taskModel BClaude Sonnet 5Anthropic57.7/1006 independent benchmarks$1.87 per solved task
Shared results
115
Claude Opus 4.8 only
15
Claude Sonnet 5 only
6
Like-for-like categories
8 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

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 65.0 from 6 benchmarks, Claude Sonnet 5 57.7 from 6 benchmarks.
Widest category
Math, 9.8 points apart on average over 6 shared benchmarks.
Widest single gap
Design Arena: spreadsheets: Claude Opus 4.8 1303 against Claude Sonnet 5 1035, 33.5 points apart on a 0 to 100 scale.
Output price
Claude Opus 4.8 $25.00 against Claude Sonnet 5 $10.00 per million output tokens, 2.5x apart.
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 Sonnet 5Ties
Coding191
Intelligence20
Agentic work1361 tie
Reasoning140
Math51
Knowledge731 tie
Web design27
General312
All shared boards93202 ties
CodingLike-for-likeSix of these boards feed the Vibe Coding Index191
Coding benchmarks for Claude Opus 4.8 and Claude Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 5Leader
LiveBench coding77.3%74.7%Claude Opus 4.8
LMArena WebDev15561540Claude Opus 4.8
AA Coding Index74.371.5Claude Opus 4.8
LiveCodeBench (Vals AI)87.8%82.4%Claude Opus 4.8
SWE-bench Verified (Vals AI)88.6%79.6%Claude Opus 4.8
Terminal-Bench 4 (Vals AI)23.2%10.6%Claude Opus 4.8
Vibe Code Bench (Vals AI)82.7%81.3%Claude Opus 4.8
Code migration (Vals AI)47.3%44.4%Claude Opus 4.8
ProgramBench (Vals AI)1.0%0.0%Claude Opus 4.8
IOI (Vals AI)Not measured45.0%Only Claude Sonnet 5 measured
Terminal-Bench 2.1 (Vals AI)71.9%74.5%Claude Sonnet 5
Vibe Code Bench 1-100 (Vals AI)Not measured13.8%Only Claude Sonnet 5 measured
SkillsBench (Vals AI)59.2%46.5%Claude Opus 4.8
CyberBench (Vals AI)Not measured70.3%Only Claude Sonnet 5 measured
SciCode54.4%54.3%Claude Opus 4.8
Terminal-Bench Hard (Artificial Analysis run)58.3%Not measuredOnly Claude Opus 4.8 measured
FrontierCode46.5%42.7%Claude Opus 4.8
GSO-Bench47.1%37.3%Claude Opus 4.8
WeirdML82.9%68.8%Claude Opus 4.8
ALE-BenchNot measured1463.1Only Claude Sonnet 5 measured
LMArena Text: coding14901483Claude Opus 4.8
LMArena Text: software and IT14881474Claude Opus 4.8
LMArena WebDev: React15571541Claude Opus 4.8
LMArena WebDev: HTML15431527Claude Opus 4.8
LMArena WebDev: from an imageNot measured1540Only Claude Sonnet 5 measured
SciCode12.3%10.7%Claude Opus 4.8
Neither model has a result on the other 13 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Claude Opus 4.8 and Claude Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 5Leader
AA Intelligence Index41.838.2Claude Opus 4.8
Epoch Capabilities Index159.0156.9Claude Opus 4.8
Agentic workLike-for-likePublic boards, shown as published, kept out of the index136
Agentic work benchmarks for Claude Opus 4.8 and Claude Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 5Leader
AA Agentic Index41.943.6Claude Sonnet 5
Design Arena: agents building web apps12331245Claude Sonnet 5
Design Arena: agents building full-stack apps12491249Tie
Design Arena: agents building mobile apps12221215Claude Opus 4.8
Design Arena: agents building Android apps12331227Claude Opus 4.8
Design Arena: agents building Godot games12311225Claude Opus 4.8
Design Arena: agents building games12391227Claude Opus 4.8
Design Arena: HTML slides11711168Claude Opus 4.8
Design Arena: PowerPoint slides12021191Claude Opus 4.8
Design Arena: spreadsheets13031035Claude Opus 4.8
LiveBench research81.8%82.8%Claude Sonnet 5
LiveBench browser96.0%72.0%Claude Opus 4.8
LiveBench long agentic48.3%39.0%Claude Opus 4.8
SREBench (Vals AI)Not measured0.4%Only Claude Sonnet 5 measured
Finance Agent (Vals AI)53.9%53.9%Claude Opus 4.8
Tax Agent Bench (Vals AI)30.0%29.4%Claude Opus 4.8
Legal Research Bench (Vals AI)43.8%41.8%Claude Opus 4.8
Excel Modeling Benchmark (Vals AI)69.4%66.3%Claude Opus 4.8
Terminal-Bench-Science (Vals AI)4.3%5.7%Claude Sonnet 5
Time Horizon Index (Vals AI)11.8%Not measuredOnly Claude Opus 4.8 measured
τ²-Bench Telecom94.4%Not measuredOnly Claude Opus 4.8 measured
APEX-Agents48.9%54.5%Claude Sonnet 5
Remote Labor Index8.3%Not measuredOnly Claude Opus 4.8 measured
OSWorld 2.020.6%Not measuredOnly Claude Opus 4.8 measured
DeepResearch Bench50.2%Not measuredOnly Claude Opus 4.8 measured
PostTrainBench33.8%Not measuredOnly Claude Opus 4.8 measured
Vending-Bench 2$5,787$6,378Claude Sonnet 5
LMArena Search1204Not measuredOnly Claude Opus 4.8 measured
Neither model has a result on the other 13 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index140
Reasoning benchmarks for Claude Opus 4.8 and Claude Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 5Leader
GPQA Diamond92.4%88.9%Claude Opus 4.8
LiveBench reasoning89.2%88.7%Claude Opus 4.8
MMMU Pro (Vals AI)86.6%83.0%Claude Opus 4.8
GPQA Diamond (Artificial Analysis run)92.0%91.1%Claude Opus 4.8
Humanity's Last Exam (Artificial Analysis run)48.7%41.3%Claude Opus 4.8
GPQA Diamond (Epoch AI run)91.0%90.5%Claude Opus 4.8
ARC-AGI92.5%Not measuredOnly Claude Opus 4.8 measured
ARC-AGI-272.1%Not measuredOnly Claude Opus 4.8 measured
SimpleBench64.8%60.6%Claude Opus 4.8
Chess puzzles52.0%45.0%Claude Opus 4.8
Mystery game puzzles36.0%35.0%Claude Opus 4.8
EBR-bench28.6%Not measuredOnly Claude Opus 4.8 measured
LMArena Text: hard prompts14821461Claude Opus 4.8
LMArena Text: expert15031489Claude Opus 4.8
LMArena Text: hard prompts in English14871464Claude Opus 4.8
LMArena Vision: diagrams13151293Claude Opus 4.8
LMArena Vision: homework13371311Claude Opus 4.8
Kagi LLM Benchmark88.8%Not measuredOnly Claude Opus 4.8 measured
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index51
Math benchmarks for Claude Opus 4.8 and Claude Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 5Leader
OTIS Mock AIME98.3%94.7%Claude Opus 4.8
FrontierMath, tiers 1 to 380.0%65.6%Claude Opus 4.8
FrontierMath, tier 456.1%29.3%Claude Opus 4.8
ProofBench69.0%77.0%Claude Sonnet 5
LMArena Text: math14861467Claude Opus 4.8
LMArena Text: mathematical work14971481Claude Opus 4.8
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index73
Knowledge benchmarks for Claude Opus 4.8 and Claude Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 5Leader
MMLU Pro89.6%87.5%Claude Opus 4.8
LegalBench (Vals AI)83.6%83.9%Claude Sonnet 5
CorpFin (Vals AI)66.7%67.0%Claude Sonnet 5
TaxEval (Vals AI)75.6%75.6%Tie
MortgageTax (Vals AI)69.9%70.0%Claude Sonnet 5
MedCode (Vals AI)53.2%47.5%Claude Opus 4.8
SimpleQA Verified53.0%33.7%Claude Opus 4.8
LMArena Text: science14771460Claude Opus 4.8
LMArena Text: medicine and healthcare14641442Claude Opus 4.8
LMArena Text: legal and government14771454Claude Opus 4.8
LMArena Text: business and finance14581436Claude Opus 4.8
FACTS Benchmark Suite53.5%Not measuredOnly Claude Opus 4.8 measured
Neither model has a result on the other 6 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index27
Web design benchmarks for Claude Opus 4.8 and Claude Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 5Leader
Design Arena: websites12581274Claude Sonnet 5
Design Arena: UI components12551277Claude Sonnet 5
Design Arena: game dev12641277Claude Sonnet 5
Design Arena: data viz12401251Claude Sonnet 5
Design Arena: 3D12221261Claude Sonnet 5
Design Arena: all code categories12551275Claude Sonnet 5
Design Arena: SVG11951192Claude Opus 4.8
Design Arena: ASCII art12771207Claude Opus 4.8
LiveBench design78.8%82.0%Claude Sonnet 5
GeneralLike-for-likePublic boards, shown as published, kept out of the index312
General benchmarks for Claude Opus 4.8 and Claude Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 5Leader
LiveBench data analysis66.0%71.7%Claude Sonnet 5
Public Benefits Bench (Vals AI)68.1%66.0%Claude Opus 4.8
SAGE (Vals AI)54.8%48.9%Claude Opus 4.8
MedScribe (Vals AI)85.8%76.0%Claude Opus 4.8
Vals Multimodal Index70.9%68.8%Claude Opus 4.8
IFBench62.2%Not measuredOnly Claude Opus 4.8 measured
Long context reasoning (AA-LCR)77.7%82.0%Claude Sonnet 5
DTBench94.9%92.5%Claude Opus 4.8
LMCA57.5%50.0%Claude Opus 4.8
Surface Evolver Bench87.5%60.0%Claude Opus 4.8
Furniture assembly42.5%Not measuredOnly Claude Opus 4.8 measured
LMArena Text14611442Claude Opus 4.8
LMArena Text: instruction following14761452Claude Opus 4.8
LMArena Text: creative writing14541416Claude Opus 4.8
LMArena Vision12941273Claude Opus 4.8
LMArena Text: English14681450Claude Opus 4.8
LMArena Text: non-English14501431Claude Opus 4.8
LMArena Text: Chinese15061478Claude Opus 4.8
LMArena Text: German14721442Claude Opus 4.8
LMArena Text: Japanese14421422Claude Opus 4.8
LMArena Text: Korean14311412Claude Opus 4.8
LMArena Text: French14811463Claude Opus 4.8
LMArena Text: Spanish14651438Claude Opus 4.8
LMArena Text: Russian14741450Claude Opus 4.8
LMArena Text: Polish14591425Claude Opus 4.8
LMArena Text: multi-turn14761453Claude Opus 4.8
LMArena Text: longer queries14831462Claude Opus 4.8
LMArena Text: writing and language14611436Claude Opus 4.8
LMArena Text: entertainment and media14401411Claude Opus 4.8
LMArena Vision: English12931275Claude Opus 4.8
LMArena Vision: Chinese13461331Claude Opus 4.8
LMArena Vision: reading text13031288Claude Opus 4.8
LMArena Vision: humor12821261Claude Opus 4.8
LMArena Vision: creative writing13011246Claude Opus 4.8
LMArena Document14751466Claude Opus 4.8
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 Sonnet 5
BenchmarkClaude Opus 4.8Claude Sonnet 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 Sonnet 5$0.0070Fits in one request

Claude Sonnet 5 costs less

Repository review

50K input + 3K output tokens

Claude Opus 4.8$0.325Fits in one request
Claude Sonnet 5$0.130Fits in one request

Claude Sonnet 5 costs less

Agent loop

200K input + 10K output tokens

Claude Opus 4.8$1.25Fits in one request
Claude Sonnet 5$0.500Fits in one request

Claude Sonnet 5 costs less

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 Sonnet 5
SpecificationAnthropic Claude Opus 4.8Anthropic Claude Sonnet 5
MakerAnthropicAnthropic
WeightsNo public weights listedNo public weights listed
ReleasedMay 27, 2026June 30, 2026
Context window1M1M
Input $/Mtok$5.00$2.00
Output $/Mtok$25.00$10.00
Coding index65.057.7
Independent measurements behind it66
Cost per solved task$2.17$1.87
SourceOpenRouter listingOpenRouter listing

Questions

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

Pick Claude Opus 4.8 for coding: it scores 65.0 on the Vibe Coding Index against 57.7 for Claude Sonnet 5. Pick Claude Sonnet 5 for output tokens 2.5x cheaper ($10.00 vs $25.00 per million) and a lower cost per solved task ($1.87 against $2.17).

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

Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; Claude Sonnet 5 costs $2.00 and $10.00. Per solved SWE-bench Verified task, Claude Opus 4.8 costs $2.17 and Claude Sonnet 5 $1.87.

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

Neither: both take 1M tokens.

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

115 benchmarks have a result for both models as of October 22, 2026. Claude Opus 4.8's coding index rests on 6 independent benchmarks and Claude Sonnet 5's on 6. 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-sonnet-5" title="Claude Opus 4.8 vs Claude Sonnet 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-sonnet-5">LLMBenchmark.ai, Claude Opus 4.8 vs Claude Sonnet 5</a></p>