vibecodingnav
中/EN
Models/gemini-3-8-flash

Gemini 3.8 Flash

11th on sudo L7 at 10.6% (High). Same score as Grok 4.7, at the back of the public board.

Google · Proprietary · Fetched 2026-10-10

LMArena WebDev
1583
#31 / 142 · ±8
LMArena Image-to-WebDev
1593
#19 / 64 · ±14
Android Bench
8.0%
#8 / 10 · 3.3–13.3

Benchmark scores

Boards older than 90 days are not counted as recommendations
BenchmarkSettingScoreRankCostAs of
LMArena WebDevhigh1583±8#31—2026-10-08
LMArena Image-to-WebDevhigh1593±14#19—2026-10-09
Android Benchantigravity-sdk8.03.3–13.3#8$34.52026-10-06
Terminal-Bench 4.0high · mini-SWE-agent19.1±3.4#29$1.8k2026-10-06
The Slop Index—48#35—2026-10-11
sudo L7High reasoning10.6#12—2026-10-09
Next.js Agent EvalsOpenCode90.0#6$0.6052026-09-25
Convex Evals1 run64.29#10$2.402026-10-08
ReactBenchmax34.430.3–38.4#23$1.632026-10-10
DeepSWE v1.1high · mini-swe-agent73.872.4–75.2#2$2.362026-09-22
SWE-bench Verified—This board has stopped updating——2026-02-26
Aider Polyglot—This board has stopped updating——2025-11-20

Use it or not

Why
  • The official sudo L7 board includes this tier.
When not
  • 10.6% is far from the 45% leaders. A poor default for production judgment.

Appears in these scenarios

Collected from scenario boards
Build a Web frontend#2DeepSWE 73.8%Build an Android app#2DeepSWE 73.8%Write CLI, scripts, and DevOps#2DeepSWE 73.8%Write docs & copy#35Slop Index 48

Peers

ModelWebDevPrice $/M
Gemini 3.8 FlashCurrent1583—Claude Haiku 5.51587—