Compare / Gemini 3.5 Flash vs Gemini 3.8 Flash

Gemini 3.5 Flash vs Gemini 3.8 Flash

Which is cheaper, which is more capable, and which does better at each kind of work, from 33 results on 8 sources that measured both models.

GoogleGemini 3.5 Flash Released 19 May 2026 · $1.50 in · $9.00 out
GoogleGemini 3.8 Flash Released 2 Sep 2026 · $1.50 in · $7.50 out
Gemini 3.5 Flash wins 85 too close to callGemini 3.8 Flash wins 20

Of the 33 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).

Gemini 3.5 Flash or Gemini 3.8 Flash? The short answer

Pick Gemini 3.8 Flash: it scores higher (Intelligence Index 40.9 against 33.6) and costs less.

  • Gemini 3.8 Flash is cheaper to run: $0.015 against $0.029 per game of bullet chess on BulletBench (47% less).
  • Gemini 3.8 Flash scores higher on the Artificial Analysis Intelligence Index: 40.9 against 33.6.
  • Gemini 3.8 Flash is newer: released 2 Sep 2026, against 19 May 2026 for Gemini 3.5 Flash.
  • On our own benchmarks, on the results both models have: Gemini 3.5 Flash does better for speed under pressure.
  • On third-party benchmarks, on the results both models have: Gemini 3.8 Flash does better for agents and tool use, professional work, reasoning and knowledge, writing and sticking to the facts.

Head to head

33.6
Intelligence Index
40.9
$3.38
Price per million tokens, blendedlower is better
$3.00
1,048,576
Context window, tokens
1,048,576

Every fact side by side

Gemini 3.5 FlashGemini 3.8 Flash
Price per million tokens$1.50 in · $9.00 out$1.50 in · $7.50 out
Blended (3 in : 1 out)$3.38$3.00
Cheapest host, blended$3.38 (Google)$1.50 (Google)
Cost per game of bullet chess (BulletBench)$0.029$0.015
Cached input, per million$0.15 cached input–
Intelligence Index33.640.9
Spring Prompt overall–58
Output speed, tokens a second–141
Context window1,048,576 tokens1,048,576 tokens
Released19 May 20262 Sep 2026 (newer)
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 not measured 22 of 32
Decks from an analysisTurning a finished analysis into a deck you could present as it is not measured 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 not measured 11 of 32
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls 49 of 159 35 of 159
Professional workReal tasks from banking, consulting and law, business documents and freelance projects 42 of 51 11 of 51
Reasoning and knowledgeHard questions across science, maths and general knowledge 38 of 187 15 of 187
WritingWhat people prefer in blind comparisons, and judged writing quality 13 of 169 3 of 169
Sticking to the factsSummarising without inventing things, and factual answers 16 of 139 12 of 139
Speed under pressureGood decisions against a real clock (fast chess) 4 of 35 13 of 35

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.

MeasureGemini 3.5 FlashGemini 3.8 Flash
BulletBench measured by Spring Prompt · 3 results
Cost per game · Bullet 60sUS dollars, lower is better 51 · 43 of 59 measured $0.0289minimal reasoning $0.0154
Ladder Elo · Bullet 60sladder Elo, higher is better 5 · 1 of 59 measured 810minimal reasoning 1,068low reasoning
Ladder Elo · Lightning 10+1ladder Elo, higher is better 4 · 40 of 45 measured 727minimal reasoning 197low reasoning
APEX-Agents reported by APEX-Agents · 1 result
Tasks passed% of tasks, higher is better 35 · 8 of 37 measured 27.5% 64.3%
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results
Overall · TextArena rating, higher is better 25 · 7 of 387 measured 1,477high reasoning 1,495high reasoning
Overall · Text factualityArena rating, higher is better 17 · 16 of 163 measured 1,476high reasoning 1,477high reasoning
Business, management and finance · TextArena rating, higher is better 35 · 21 of 380 measured 1,468high reasoning 1,481high reasoning
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 49 · 25 of 249 measured 33.6medium reasoning 40.9high reasoning
GPQA Diamond% of questions, higher is better 23 · 2 of 232 measured 92.2%high reasoning 95.3%high reasoning
Humanity's Last Exam% of questions, higher is better 34 · 18 of 250 measured 42.7%high reasoning 47.8%high reasoning
GDP.pdf reported by GDP.pdf · 1 result
Professional document tasks% of rubric, higher is better 24 · 12 of 32 measured 14.0% 23.4%medium reasoning
SimpleQA Verified (Epoch AI) reported by SimpleQA Verified (Epoch AI) · 1 result
Correct answers% of questions, higher is better 11 · 7 of 76 measured 66.2%high reasoning 69.7%high reasoning
UGI Leaderboard reported by UGI Leaderboard · 3 results
Writing score · UGIscore out of 100, higher is better 4 · 1 of 230 measured 72.5medium reasoning 78.6medium reasoning
Requested-length error · UGI% off the requested word count, lower is better 33 · 57 of 230 measured 9.0%low reasoning 13.0%high reasoning
Style adherence · UGIscore from 0 to 1, higher is better 96 · 132 of 230 measured 0.35medium reasoning 0.34high reasoning
Vending-Bench 2 reported by Vending-Bench 2 · 1 result
Money after a yearUS dollars, higher is better 23 · 26 of 58 measured $5,396 $5,094
Show all 33 results (17 more)
MeasureGemini 3.5 FlashGemini 3.8 Flash
BulletBench measured by Spring Prompt · 7 results
Cost per game · Lightning 10+1US dollars, lower is better 42 · 30 of 45 measured $0.0213minimal reasoning $0.0041low reasoning
Games lost on time · Bullet 60s% of games, lower is better 31 · 39 of 59 measured 25.0%minimal reasoning 50.0%low reasoning
Games lost on time · Lightning 10+1% of games, lower is better 34 · 42 of 45 measured 33.3%minimal reasoning 83.3%low reasoning
Invalid moves · Bullet 60s% of moves, lower is better 56 · 1 of 59 measured 0.5%minimal reasoning 0.0%
Invalid moves · Lightning 10+1% of moves, lower is better 40 · 1 of 45 measured 0.6%minimal reasoning 0.0%low reasoning
Median move time · Bullet 60smilliseconds, lower is better 35 · 46 of 59 measured 1.2 sminimal reasoning 1.8 slow reasoning
Median move time · Lightning 10+1milliseconds, lower is better 38 · 44 of 45 measured 1.2 sminimal reasoning 1.8 slow reasoning
Arena (formerly LMArena) reported by Arena (formerly LMArena) · 4 results
Creative writing · TextArena rating, higher is better 15 · 7 of 385 measured 1,469medium reasoning 1,485high reasoning
Expert prompts · TextArena rating, higher is better 49 · 13 of 338 measured 1,492high reasoning 1,526high reasoning
Instruction following · TextArena rating, higher is better 35 · 12 of 387 measured 1,466medium reasoning 1,486high reasoning
Writing, literature and language · TextArena rating, higher is better 15 · 8 of 386 measured 1,474high reasoning 1,487high reasoning
Artificial Analysis reported by Artificial Analysis · 6 results
Terminal-Bench 2.1% of tasks, higher is better 26 · 8 of 131 measured 78.7%high reasoning 87.6%high reasoning
Terminal-Bench 4.0% of tasks, higher is better 48 · 30 of 120 measured 6.6%high reasoning 19.7%high reasoning
τ-bench banking% of tasks, higher is better 31 · 11 of 121 measured 32.2%high reasoning 45.8%medium reasoning
Artificial Analysis Coding Indexindex score, higher is better 27 · 9 of 130 measured 70.1high reasoning 76.3high reasoning
Long-context reasoning (AA-LCR)% of questions, higher is better 89 · 7 of 241 measured 74.3%medium reasoning 84.0%medium reasoning
SciCode% of problems, higher is better 36 · 20 of 121 measured 53.9%high reasoning 56.6%high reasoning

Quick answers

Is Gemini 3.5 Flash better than Gemini 3.8 Flash?

Pick Gemini 3.8 Flash: it scores higher (Intelligence Index 40.9 against 33.6) and costs less.

Which is cheaper, Gemini 3.5 Flash or Gemini 3.8 Flash?

Gemini 3.8 Flash is cheaper to run: $0.015 against $0.029 per game of bullet chess on BulletBench (47% less).

Which is better for agents and tool use, Gemini 3.5 Flash or Gemini 3.8 Flash?

Gemini 3.8 Flash ranks 35 of 159 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 49 for Gemini 3.5 Flash.

Which is better for professional work, Gemini 3.5 Flash or Gemini 3.8 Flash?

Gemini 3.8 Flash ranks 11 of 51 models on sale for real tasks from banking, consulting and law, business documents and freelance projects, against 42 for Gemini 3.5 Flash.

Which is better for reasoning and knowledge, Gemini 3.5 Flash or Gemini 3.8 Flash?

Gemini 3.8 Flash ranks 15 of 187 models on sale for hard questions across science, maths and general knowledge, against 38 for Gemini 3.5 Flash.

Which is better for writing, Gemini 3.5 Flash or Gemini 3.8 Flash?

Gemini 3.8 Flash ranks 3 of 169 models on sale for what people prefer in blind comparisons, and judged writing quality, against 13 for Gemini 3.5 Flash.

Which is better for sticking to the facts, Gemini 3.5 Flash or Gemini 3.8 Flash?

Gemini 3.8 Flash ranks 12 of 139 models on sale for summarising without inventing things, and factual answers, against 16 for Gemini 3.5 Flash.

Run Gemini 3.5 Flash 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.