Compare / Claude Sonnet 5.5 vs Gemini 3.8 Flash

Claude Sonnet 5.5 vs Gemini 3.8 Flash

Which is cheaper, which is more capable, and which does better at each kind of work, from 87 results on 8 sources that measured both models.

AnthropicClaude Sonnet 5.5 Released 28 Sep 2026 · $2.00 in · $10.00 out
GoogleGemini 3.8 Flash Released 2 Sep 2026 · $1.50 in · $7.50 out
Claude Sonnet 5.5 wins 3323 too close to callGemini 3.8 Flash wins 31

Of the 87 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.

Results as of 16 November 2026 (catalogue release 2026-11-16-49bdfa045dee). See also Anthropic vs Google, every model from both labs.

Claude Sonnet 5.5 or Gemini 3.8 Flash? The short answer

They score about the same, so pick Gemini 3.8 Flash: it costs 40% less per product listing on CatalogBench.

  • Spring Prompt overall: Claude Sonnet 5.5 61, Gemini 3.8 Flash 58 (out of 100; how the score works).
  • Gemini 3.8 Flash is cheaper to run: $0.010 against $0.017 per product listing on CatalogBench (40% less).
  • Claude Sonnet 5.5 scores higher on the Artificial Analysis Intelligence Index: 56.0 against 40.9.
  • Claude Sonnet 5.5 is newer: released 28 Sep 2026, against 2 Sep 2026 for Gemini 3.8 Flash.
  • On our own benchmarks, on the results both models have: Claude Sonnet 5.5 does better for decks from an analysis and marketing planning.
  • On third-party benchmarks, on the results both models have: Claude Sonnet 5.5 does better for agents and tool use, professional work and reasoning and knowledge; Gemini 3.8 Flash for writing and sticking to the facts.

Head to head

61
Spring Prompt overall
58
56.0
Intelligence Index
40.9
$4.00
Price per million tokens, blendedlower is better
$3.00
137
Output speed, tokens a second
141
1,000,000
Context window, tokens
1,048,576

Every fact side by side

Claude Sonnet 5.5Gemini 3.8 Flash
Price per million tokens$2.00 in · $10.00 out$1.50 in · $7.50 out
Blended (3 in : 1 out)$4.00$3.00
Cheapest host, blended$4.00 (Amazon Bedrock)$1.50 (Google)
Cost per product listing (CatalogBench)$0.017$0.010
Cached input, per million$0.20 cached input–
Intelligence Index56.040.9
Spring Prompt overall6158
Output speed, tokens a second137141
Context window1,000,000 tokens1,048,576 tokens
Released28 Sep 2026 (newer)2 Sep 2026
WeightsClosedClosed
Developer based inthe USthe US

Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.

Which is better for what

Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.

Product listingsTurning a sparse product feed and photos into listings that can go live 18 of 32 22 of 32
Decks from an analysisTurning a finished analysis into a deck you could present as it is 6 of 32 26 of 32
User surveysPlanning a user survey and reading its results without being misled not measured 16 of 23
Marketing planningPlanning a year of ad spend without overspending 7 of 32 11 of 32
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 2 of 159 35 of 159
Professional workReal tasks from banking, consulting and law, business documents and freelance projects 1 of 51 11 of 51
Reasoning and knowledgeHard questions across science, maths and general knowledge 2 of 187 15 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 33 of 169 3 of 169
Sticking to the factsSummarising without inventing things, and factual answers 42 of 139 12 of 139
Speed under pressureGood decisions against a real clock (fast chess) 32 of 35 13 of 35

Their slides, side by side

The same brief for both: Manchester store investment review for Hearthside Coffee, an invented company. First slide of each deck from DeckBench.

First slide of Claude Sonnet 5.5's deck for Hearthside Coffee
Claude Sonnet 5.5 · deck rating 1,332 · presentable as it is
First slide of Gemini 3.8 Flash's deck for Hearthside Coffee
Gemini 3.8 Flash · deck rating 788 · needs work

Every slide, side by side →

From our research

Every shared result

Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.

MeasureClaude Sonnet 5.5Gemini 3.8 Flash
BulletBench measured by Spring Prompt · 3 results
Cost per game · Blitz 3+2US dollars, lower is better 12 · 11 of 20 measured $0.12 $0.0795
Ladder Elo · Lightning 10+1ladder Elo, higher is better 42 · 40 of 45 measured 33low reasoning 197low reasoning
Cost per game · Lightning 10+1US dollars, lower is better 38 · 30 of 45 measured $0.0080low reasoning $0.0041low reasoning
CatalogBench measured by Spring Prompt · 3 results
Cost per productUS dollars, lower is better 21 · 18 of 33 measured $0.0172 $0.0103
Reliably publish-ready% of products, higher is better 21 · 20 of 32 measured 32.1% 35.7%
Reliably publish-ready · sales brief% of products, higher is better 17 · 23 of 32 measured 10.7% 3.6%
DeckBench measured by Spring Prompt · 3 results
Cost per deckUS dollars, lower is better 20 · 15 of 33 measured $0.15 $0.0613
Deck ratingrating, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 6 · 26 of 33 measured 1,332 788
Accurate decks% of tasks, higher is better 9 · 24 of 33 measured 66.7% 16.7%
ROASBench measured by Spring Prompt · 3 results
Overall scorescore out of 100, higher is better 7 · 11 of 33 measured 52.0 51.4
Audience scorescore out of 100, higher is better 18 · 10 of 33 measured 61.8 67.1
Business scorescore out of 100, higher is better 5 · 11 of 33 measured 57.2 51.0
APEX-Agents reported by APEX-Agents · 1 result
Tasks passed% of tasks, higher is better 1 · 8 of 37 measured 75.5%max reasoning 64.3%
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 39 · 7 of 387 measured 1,471xhigh reasoning 1,495high reasoning
Overall · Text factualityArena rating, higher is better 24 · 16 of 163 measured 1,470xhigh reasoning 1,477high reasoning
Business, management and finance · TextArena rating, higher is better 46 · 21 of 380 measured 1,464xhigh reasoning 1,481high reasoning
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 2 · 25 of 249 measured 56.0max reasoning 40.9high reasoning
Humanity's Last Exam% of questions, higher is better 5 · 18 of 250 measured 55.0%max reasoning 47.8%high reasoning
Terminal-Bench 4.0% of tasks, higher is better 2 · 30 of 120 measured 63.6%max reasoning 19.7%high reasoning
SimpleQA Verified (Epoch AI) reported by SimpleQA Verified (Epoch AI) · 1 result
Correct answers% of questions, higher is better 33 · 7 of 76 measured 46.5%max reasoning 69.7%high reasoning
Show all 87 results (67 more)
MeasureClaude Sonnet 5.5Gemini 3.8 Flash
BulletBench measured by Spring Prompt · 7 results
Games lost on time · Blitz 3+2% of games, lower is better 9 · 7 of 20 measured 37.5% 12.5%
Games lost on time · Lightning 10+1% of games, lower is better 44 · 42 of 45 measured 91.7%low reasoning 83.3%low reasoning
Invalid moves · Blitz 3+2% of moves, lower is better 1 · 1 of 20 measured 0.0% 0.0%
Invalid moves · Lightning 10+1% of moves, lower is better 1 · 1 of 45 measured 0.0%low reasoning 0.0%low reasoning
Ladder Elo · Blitz 3+2ladder Elo, higher is better 11 · 2 of 20 measured 557 1,126
Median move time · Blitz 3+2milliseconds, lower is better 19 · 9 of 20 measured 8.1 s 4.0 s
Median move time · Lightning 10+1milliseconds, lower is better 45 · 44 of 45 measured 2.1 slow reasoning 1.8 slow reasoning
CatalogBench measured by Spring Prompt · 28 results
Channel compliance% of products, higher is better 25 · 25 of 33 measured 99.4% 99.4%
Channel rules broken% of products, lower is better 27 · 20 of 33 measured 1.2% 0.6%
Channel rules broken · sales brief% of products, lower is better 17 · 17 of 33 measured 0.6% 0.6%
Claims to check% of products, lower is better 1 · 1 of 31 measured 0.0% 0.0%
Claims to check · sales brief% of products, lower is better 19 · 26 of 31 measured 0.6% 5.4%
Conflicts caught% of conflicts, higher is better 1 · 1 of 33 measured 100.0% 100.0%
Content quality% of checks, higher is better 22 · 9 of 33 measured 96.5% 97.7%
Content quality · sales brief% of checks, higher is better 16 · 14 of 33 measured 97.3% 97.6%
Decision accuracy% of decisions, higher is better 24 · 19 of 33 measured 95.7% 96.4%
Failed outputs% of products, lower is better 1 · 1 of 33 measured 0.0% 0.0%
Failed outputs · sales brief% of products, lower is better 1 · 1 of 33 measured 0.0% 0.0%
Field accuracy% of missing fields, higher is better 30 · 24 of 33 measured 91.5% 92.7%
Invented values% of filled values, lower is better 9 · 4 of 33 measured 0.8% 0.3%
Missing UK information% of products, lower is better 1 · 1 of 33 measured 0.0% 0.0%
Missing UK information · sales brief% of products, lower is better 1 · 1 of 33 measured 0.0% 0.0%
Not findable% of products, lower is better 4 · 27 of 33 measured 1.2% 8.3%
Not findable · sales brief% of products, lower is better 4 · 28 of 33 measured 0.6% 8.9%
Publish-ready listings% of products, higher is better 24 · 21 of 33 measured 46.4% 54.8%
Publish-ready listings · sales brief% of products, higher is better 19 · 23 of 33 measured 27.4% 16.7%
Time per productseconds, lower is better 9 · 19 of 33 measured 9.90 18.7
Unsupported claims% of products, lower is better 28 · 24 of 33 measured 39.3% 31.6%
Unsupported claimsclaims per product, lower is better 25 · 22 of 33 measured 1.08 0.79
Unsupported claims · sales brief% of products, lower is better 17 · 28 of 33 measured 59.5% 81.6%
Unsupported claims · sales briefclaims per product, lower is better 17 · 24 of 33 measured 1.71 3.08
Wrong attributes% of products, lower is better 22 · 11 of 33 measured 19.1% 15.5%
Wrong attributes · sales brief% of products, lower is better 25 · 2 of 33 measured 20.8% 11.9%
Wrong category or variant% of products, lower is better 24 · 27 of 33 measured 3.0% 3.6%
Wrong category or variant · sales brief% of products, lower is better 23 · 23 of 33 measured 3.0% 3.0%
DeckBench measured by Spring Prompt · 14 results
Caveat dropped% of tasks, lower is better 15 · 1 of 33 measured 16.7% 0.0%
Clean layout% of tasks, higher is better 1 · 24 of 33 measured 100.0% 16.7%
Design quality% of the maximum, higher is better 7 · 19 of 33 measured 78.5% 61.7%
Draft figure quoted% of tasks, lower is better 1 · 26 of 33 measured 0.0% 50.0%
Findings missing% of tasks, lower is better 1 · 1 of 33 measured 0.0% 0.0%
Head-to-head win rate% of comparisons, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 4 · 26 of 33 measured 82.2% 38.7%
Layout defects% of tasks, lower is better 1 · 24 of 33 measured 0.0% 83.3%
Misleading metric used% of tasks, lower is better 1 · 1 of 33 measured 0.0% 0.0%
Presentable decks% of tasks, higher is better 3 · 15 of 33 measured 50.0% 0.0%
Recommendation late or wrong% of tasks, lower is better 15 · 20 of 33 measured 16.7% 33.3%
Slides needing work% of tasks, lower is better 10 · 10 of 33 measured 33.3% 33.3%
Time per deckseconds, lower is better 11 · 13 of 33 measured 83.4 93.0
Unsupported claims% of tasks, lower is better 1 · 30 of 33 measured 0.0% 50.0%
Unsupported numbers% of tasks, lower is better 1 · 26 of 33 measured 0.0% 16.7%
ROASBench measured by Spring Prompt · 6 results
Consistency scorescore out of 100, higher is better 17 · 6 of 33 measured 28.5 36.4
Contribution profitsimulated US dollars, higher is better 20 · 5 of 33 measured $571,203 $842,432
Cost of a runUS dollars, lower is better 20 · 15 of 33 measured $0.42 $0.16
Months over budgetmonths of 12, lower is better 1 · 1 of 33 measured 0 0
Planning scorescore out of 100, higher is better 3 · 21 of 33 measured 55.1 54.7
Return on ad spendprofit per $1 spent, higher is better 4 · 12 of 33 measured $0.70 $0.65
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 8 results
Confirmed task success · AgentIPS effect estimate, higher is better 5 · 11 of 48 measured 0.14max reasoning, arena agent 0.07high reasoning, arena agent
Creative writing · TextArena rating, higher is better 34 · 7 of 385 measured 1,447xhigh reasoning 1,485high reasoning
Expert prompts · TextArena rating, higher is better 1 · 13 of 338 measured 1,556xhigh reasoning 1,526high reasoning
Instruction following · TextArena rating, higher is better 15 · 12 of 387 measured 1,484xhigh reasoning 1,486high reasoning
Praise over complaint · AgentIPS effect estimate, higher is better 5 · 19 of 48 measured 0.26max reasoning, arena agent 0.05high reasoning, arena agent
Steerability · AgentIPS effect estimate, higher is better 8 · 17 of 48 measured 0.07max reasoning, arena agent 0.03high reasoning, arena agent
Tool grounding · AgentIPS effect estimate, higher is better 7 · 17 of 48 measured 0.00max reasoning, arena agent 0.00high reasoning, arena agent
Writing, literature and language · TextArena rating, higher is better 27 · 8 of 386 measured 1,461xhigh reasoning 1,487high reasoning
Artificial Analysis reported by Artificial Analysis · 4 results
Long-context reasoning (AA-LCR)% of questions, higher is better 21 · 7 of 241 measured 82.7%max reasoning 84.0%medium reasoning
Output speedtokens per second, higher is better 33 · 31 of 78 measured 137max reasoning 141high reasoning
SciCode% of problems, higher is better 4 · 20 of 121 measured 61.0%max reasoning 56.6%high reasoning
Time to first answer tokenseconds, lower is better 17 · 58 of 78 measured 0.71low reasoning 15.5high reasoning

Quick answers

Is Claude Sonnet 5.5 better than Gemini 3.8 Flash?

They score about the same, so pick Gemini 3.8 Flash: it costs 40% less per product listing on CatalogBench.

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

Gemini 3.8 Flash is cheaper to run: $0.010 against $0.017 per product listing on CatalogBench (40% less).

Which is faster, Claude Sonnet 5.5 or Gemini 3.8 Flash?

Gemini 3.8 Flash writes about 141 tokens a second on its usual API, against 137 for Claude Sonnet 5.5 (Artificial Analysis).

Which is better for product listings, Claude Sonnet 5.5 or Gemini 3.8 Flash?

Claude Sonnet 5.5 ranks 18 of 32 models on sale for turning a sparse product feed and photos into listings that can go live, against 22 for Gemini 3.8 Flash.

Which is better for decks from an analysis, Claude Sonnet 5.5 or Gemini 3.8 Flash?

Claude Sonnet 5.5 ranks 6 of 32 models on sale for turning a finished analysis into a deck you could present as it is, against 26 for Gemini 3.8 Flash.

Which is better for marketing planning, Claude Sonnet 5.5 or Gemini 3.8 Flash?

Claude Sonnet 5.5 ranks 7 of 32 models on sale for planning a year of ad spend without overspending, against 11 for Gemini 3.8 Flash.

Which is better for agents and tool use, Claude Sonnet 5.5 or Gemini 3.8 Flash?

Claude Sonnet 5.5 ranks 2 of 159 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 35 for Gemini 3.8 Flash.

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

Gemini 3.8 Flash: 1,048,576 tokens, against 1,000,000 for Claude Sonnet 5.5.

Run Claude Sonnet 5.5 and Gemini 3.8 Flash on your own prompt.

Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.