LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Claude Opus 5.5 vs Gemini 3.8 Flash

Updated October 21, 2026. 102 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Claude Opus 5.5 Replace Gemini 3.8 Flash
Claude Opus 5.5Anthropic
Coding
92.2
Overall
57.6
Output $/M
$20.00
92
Gemini 3.8 FlashGoogle
Coding
70.5
Overall
40.9
Output $/M
$4.125
9
vs boards won of 102 shared, 1 tied

Decision reading

Pick Claude Opus 5.5 for coding: it scores 92.2 on the Vibe Coding Index against 70.5 for Gemini 3.8 Flash.

Pick Gemini 3.8 Flash when output tokens cost 4.8x less, or when the context window is larger, 1.05M against 1M tokens.

Model AClaude Opus 5.5Anthropic92.2/1002 independent benchmarksNo cost per solved taskModel BGemini 3.8 FlashGoogle70.5/1002 independent benchmarks$2.74 per solved task
Shared results
102
Claude Opus 5.5 only
5
Gemini 3.8 Flash only
24
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 5.5 92.2 from 2 benchmarks, Gemini 3.8 Flash 70.5 from 2 benchmarks.
Widest category
Math, 25.3 points apart on average over 6 shared benchmarks.
Widest single gap
FrontierMath, tier 4: Claude Opus 5.5 95.0% against Gemini 3.8 Flash 22.0%, 73.0 points apart on a 0 to 100 scale.
Output price
Claude Opus 5.5 $20.00 against Gemini 3.8 Flash $4.125 per million output tokens, 4.8x apart.
Context window
Claude Opus 5.5 takes 1M tokens, Gemini 3.8 Flash 1.05M.

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 5.5BarsGemini 3.8 FlashTies
Coding200
Intelligence20
Agentic work182
Reasoning1111 tie
Math42
Knowledge60
Web design80
General234
All shared boards9291 tie
CodingLike-for-likeFour of these boards feed the Vibe Coding Index200
Coding benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
LiveBench coding96.3%82.2%Claude Opus 5.5
LMArena WebDev18071585Claude Opus 5.5
AA Coding IndexNot measured76.3Only Gemini 3.8 Flash measured
LiveCodeBench (Vals AI)Not measured89.5%Only Gemini 3.8 Flash measured
SWE-bench Verified (Vals AI)Not measured80.0%Only Gemini 3.8 Flash measured
Terminal-Bench 4 (Vals AI)65.2%19.7%Claude Opus 5.5
Vibe Code Bench (Vals AI)90.3%78.7%Claude Opus 5.5
Code migration (Vals AI)66.7%36.5%Claude Opus 5.5
ProgramBench (Vals AI)18.5%1.0%Claude Opus 5.5
IOI (Vals AI)95.1%56.9%Claude Opus 5.5
Terminal-Bench 2.1 (Vals AI)87.6%81.3%Claude Opus 5.5
SWE-bench Multimodal (Vals AI)70.2%40.8%Claude Opus 5.5
Vibe Code Bench 1-100 (Vals AI)30.4%18.8%Claude Opus 5.5
SkillsBench (Vals AI)Not measured58.0%Only Gemini 3.8 Flash measured
CyberBench (Vals AI)74.6%43.8%Claude Opus 5.5
SciCode66.9%56.6%Claude Opus 5.5
FrontierCode54.6%41.2%Claude Opus 5.5
FrontierSWE62.3%19.6%Claude Opus 5.5
WeirdMLNot measured84.8%Only Gemini 3.8 Flash measured
MirrorCode77.4%Not measuredOnly Claude Opus 5.5 measured
ALE-Bench2147.41269.7Claude Opus 5.5
LMArena Text: coding15451508Claude Opus 5.5
LMArena Text: software and IT15381507Claude Opus 5.5
LMArena WebDev: React18551577Claude Opus 5.5
LMArena WebDev: HTML17531616Claude Opus 5.5
LMArena WebDev: from an image17531593Claude Opus 5.5
Neither model has a result on the other 12 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
AA Intelligence Index57.640.9Claude Opus 5.5
Epoch Capabilities Index168.8157.4Claude Opus 5.5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index182
Agentic work benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
AA Agentic IndexNot measured40.2Only Gemini 3.8 Flash measured
Design Arena: agents building web apps13161238Claude Opus 5.5
Design Arena: agents building full-stack appsNot measured1244Only Gemini 3.8 Flash measured
Design Arena: agents building mobile apps13111236Claude Opus 5.5
Design Arena: agents building Android appsNot measured1078Only Gemini 3.8 Flash measured
Design Arena: agents building Godot gamesNot measured1273Only Gemini 3.8 Flash measured
Design Arena: PowerPoint slidesNot measured1178Only Gemini 3.8 Flash measured
Design Arena: spreadsheets13591052Claude Opus 5.5
LiveBench research85.0%79.1%Claude Opus 5.5
LiveBench browser80.0%64.0%Claude Opus 5.5
LiveBench long agentic85.0%30.1%Claude Opus 5.5
SREBench (Vals AI)33.6%11.1%Claude Opus 5.5
Finance Agent (Vals AI)58.6%61.4%Gemini 3.8 Flash
Tax Agent Bench (Vals AI)45.6%32.1%Claude Opus 5.5
Legal Research Bench (Vals AI)50.5%38.9%Claude Opus 5.5
Harvey's Legal Agent Benchmark (Vals AI)9.2%12.1%Gemini 3.8 Flash
Excel Modeling Benchmark (Vals AI)75.9%72.2%Claude Opus 5.5
BioMysteryBench (Vals AI)79.3%62.2%Claude Opus 5.5
Terminal-Bench-Science (Vals AI)47.1%8.6%Claude Opus 5.5
Vals Science Index58.7%35.8%Claude Opus 5.5
Time Horizon Index (Vals AI)91.3%18.8%Claude Opus 5.5
CUA-bench (Vals AI)14.0%4.2%Claude Opus 5.5
Robo-bench (Vals AI)14.6%Not measuredOnly Claude Opus 5.5 measured
APEX-Agents73.5%64.3%Claude Opus 5.5
Remote Labor Index21.3%5.8%Claude Opus 5.5
SpireClimb3.3%Not measuredOnly Claude Opus 5.5 measured
Vending-Bench 2$9,235$5,094Claude Opus 5.5
LMArena Search1213Not measuredOnly Claude Opus 5.5 measured
Neither model has a result on the other 11 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index111
Reasoning benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
GPQA DiamondNot measured94.4%Only Gemini 3.8 Flash measured
LiveBench reasoning92.2%89.3%Claude Opus 5.5
MysteryMechanism (Vals AI)49.5%36.5%Claude Opus 5.5
MMMU Pro (Vals AI)Not measured89.1%Only Gemini 3.8 Flash measured
GPQA Diamond (Artificial Analysis run)Not measured95.3%Only Gemini 3.8 Flash measured
Humanity's Last Exam (Artificial Analysis run)61.4%47.8%Claude Opus 5.5
GPQA Diamond (Epoch AI run)90.6%95.4%Gemini 3.8 Flash
ARC-AGI98.5%98.5%Tie
ARC-AGI-293.3%89.2%Claude Opus 5.5
Chess puzzles67.0%58.0%Claude Opus 5.5
Mystery game puzzles71.0%47.0%Claude Opus 5.5
EBR-bench71.4%Not measuredOnly Claude Opus 5.5 measured
LMArena Text: hard prompts15361509Claude Opus 5.5
LMArena Text: expert15491526Claude Opus 5.5
LMArena Text: hard prompts in English15351504Claude Opus 5.5
LMArena Vision: diagrams13451345Claude Opus 5.5
LMArena Vision: homework13621351Claude Opus 5.5
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index42
Math benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
OTIS Mock AIME100.0%98.9%Claude Opus 5.5
FrontierMath, tiers 1 to 391.2%68.4%Claude Opus 5.5
FrontierMath, tier 495.0%22.0%Claude Opus 5.5
ProofBench100.0%48.0%Claude Opus 5.5
LMArena Text: math15181532Gemini 3.8 Flash
LMArena Text: mathematical work15161519Gemini 3.8 Flash
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index60
Knowledge benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
MMLU ProNot measured90.2%Only Gemini 3.8 Flash measured
LegalBench (Vals AI)Not measured87.0%Only Gemini 3.8 Flash measured
TaxEval (Vals AI)Not measured74.5%Only Gemini 3.8 Flash measured
MortgageTax (Vals AI)Not measured65.3%Only Gemini 3.8 Flash measured
MedCode (Vals AI)49.8%48.1%Claude Opus 5.5
SimpleQA Verified72.2%69.7%Claude Opus 5.5
LMArena Text: science15241523Claude Opus 5.5
LMArena Text: medicine and healthcare15091508Claude Opus 5.5
LMArena Text: legal and government15181497Claude Opus 5.5
LMArena Text: business and finance14941479Claude Opus 5.5
Neither model has a result on the other 7 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index80
Web design benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
Design Arena: websites13351301Claude Opus 5.5
Design Arena: UI components13901318Claude Opus 5.5
Design Arena: game dev14571305Claude Opus 5.5
Design Arena: data viz13971269Claude Opus 5.5
Design Arena: 3D14771295Claude Opus 5.5
Design Arena: all code categories13901304Claude Opus 5.5
Design Arena: SVG13321295Claude Opus 5.5
Design Arena: ASCII artNot measured1264Only Gemini 3.8 Flash measured
LiveBench design86.9%72.7%Claude Opus 5.5
GeneralLike-for-likePublic boards, shown as published, kept out of the index234
General benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
Vals Index70.2%56.0%Claude Opus 5.5
LiveBench data analysis80.3%54.0%Claude Opus 5.5
Public Benefits Bench (Vals AI)70.6%65.3%Claude Opus 5.5
SAGE (Vals AI)45.8%35.1%Claude Opus 5.5
MedScribe (Vals AI)91.4%84.5%Claude Opus 5.5
Long context reasoning (AA-LCR)84.7%84.0%Claude Opus 5.5
DTBench98.9%95.7%Claude Opus 5.5
LMCA68.2%52.9%Claude Opus 5.5
Surface Evolver BenchNot measured76.9%Only Gemini 3.8 Flash measured
Furniture assembly83.3%31.7%Claude Opus 5.5
LMArena Text15161499Claude Opus 5.5
LMArena Text: instruction following15371492Claude Opus 5.5
LMArena Text: creative writing15351488Claude Opus 5.5
LMArena Vision13211315Claude Opus 5.5
LMArena Text: English15191498Claude Opus 5.5
LMArena Text: non-English15071491Claude Opus 5.5
LMArena Text: Chinese15821553Claude Opus 5.5
LMArena Text: GermanNot measured1501Only Gemini 3.8 Flash measured
LMArena Text: Japanese15341507Claude Opus 5.5
LMArena Text: KoreanNot measured1464Only Gemini 3.8 Flash measured
LMArena Text: French14931496Gemini 3.8 Flash
LMArena Text: Spanish15081487Claude Opus 5.5
LMArena Text: Russian15171514Claude Opus 5.5
LMArena Text: PolishNot measured1497Only Gemini 3.8 Flash measured
LMArena Text: multi-turn15011503Gemini 3.8 Flash
LMArena Text: longer queries15321507Claude Opus 5.5
LMArena Text: writing and language15321494Claude Opus 5.5
LMArena Text: entertainment and media15151480Claude Opus 5.5
LMArena Vision: English13171318Gemini 3.8 Flash
LMArena Vision: ChineseNot measured1416Only Gemini 3.8 Flash measured
LMArena Vision: reading text13261333Gemini 3.8 Flash
LMArena Vision: humorNot measured1321Only Gemini 3.8 Flash measured
LMArena Vision: creative writing13201299Claude Opus 5.5
Neither model has a result on the other 13 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Opus 5.5 and Gemini 3.8 Flash
BenchmarkClaude Opus 5.5Gemini 3.8 FlashLeader
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 5.5$0.014Fits in one request
Gemini 3.8 Flash$0.0029Fits in one request

Gemini 3.8 Flash costs less

Repository review

50K input + 3K output tokens

Claude Opus 5.5$0.260Fits in one request
Gemini 3.8 Flash$0.054Fits in one request

Gemini 3.8 Flash costs less

Agent loop

200K input + 10K output tokens

Claude Opus 5.5$1.00Fits in one request
Gemini 3.8 Flash$0.206Fits in one request

Gemini 3.8 Flash 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 5.5 and Gemini 3.8 Flash
SpecificationAnthropic Claude Opus 5.5Google Gemini 3.8 Flash
MakerAnthropicGoogle
WeightsNo public weights listedNo public weights listed
ReleasedSeptember 22, 2026September 2, 2026
Context window1M1.05M
Input $/Mtok$4.00$0.825
Output $/Mtok$20.00$4.125
Coding index92.270.5
Independent measurements behind it22
Cost per solved taskNot computable$2.74
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Opus 5.5 better than Gemini 3.8 Flash for coding?

Pick Claude Opus 5.5 for coding: it scores 92.2 on the Vibe Coding Index against 70.5 for Gemini 3.8 Flash. Pick Gemini 3.8 Flash when output tokens cost 4.8x less, or when the context window is larger, 1.05M against 1M tokens.

Which is cheaper, Claude Opus 5.5 or Gemini 3.8 Flash?

Claude Opus 5.5 costs $4.00 per million input tokens and $20.00 per million output tokens; Gemini 3.8 Flash costs $0.825 and $4.125. Cost per solved task is known for Gemini 3.8 Flash only.

Which has the bigger context window, Claude Opus 5.5 or Gemini 3.8 Flash?

Gemini 3.8 Flash, with 1.05M tokens against 1M.

How many benchmarks is the Claude Opus 5.5 vs Gemini 3.8 Flash comparison based on?

102 benchmarks have a result for both models as of October 21, 2026. Claude Opus 5.5's coding index rests on 2 independent benchmarks and Gemini 3.8 Flash'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-5-5-vs-gemini-3-8-flash" title="Claude Opus 5.5 vs Gemini 3.8 Flash, 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-5-5-vs-gemini-3-8-flash">LLMBenchmark.ai, Claude Opus 5.5 vs Gemini 3.8 Flash</a></p>