Skip to content

Gemini 3.8 Flash

Google's most capable Flash model, aimed at long-horizon coding, agents and enterprise workflows at Flash speed and cost. It sits just behind the leaders on the LMArena text leaderboard and ties for first on LMArena Vision.

GoogleLanguage modelReleased September 2026Data checked October 18, 2026

This model is not in the current people's rankings: it is either new and still collecting votes, or ranked below the top 12. See the AI leaderboard for the models people rate highest.

On paper

Fixed-test rankings from the benchmarks tab: Speed #4, Value #5, Reasoning #10.

Strengths and weaknesses

  • Just behind the leaders on the LMArena text leaderboard, ahead of every OpenAI model
  • Tied for first on LMArena Vision and 95.3% on GPQA Diamond (Artificial Analysis)
  • Fast output (147 tokens per second) and 89.2% on ARC-AGI-2
  • Weak at agentic terminal tasks (19.1% on Terminal-Bench 4.0)
  • The current $0.75 / $3.75 per million price doubles to $1.50 / $7.50 on January 1, 2027

Best for

  • A fast everyday assistant
  • Understanding images, video and PDFs
  • Good quality at a low price

Every score we have

Each bar compares this model with the best result any model on this page has on the same test.

Compare with