Gemini 3.8 Flash: same token price, different cost per task — a real-world billing test
JuggernautCritical92 · reddit · 2026-09-04
Google says Gemini 3.8 Flash keeps the same input/output token prices as 3.7 Flash, but 3.8 may take extra reasoning steps and call tools more often on hard tasks—so a flat token price can still mean a different bill per completed task.
Google's advice: lower the effort setting or stay on 3.7 Flash when efficiency matters; planning and difficult code changes may benefit from the extra work, while classifiers and routine extraction may just burn more tokens.
The author tested both slugs through the same ZenMux API setup with identical tasks and timeouts, tracking retries and accepted patches. Both did the work well with no meaningful difference—but the author admits the tasks were too easy. The real question is whether 3.8 completes enough extra tasks to cover its extra reasoning, which requires accepted results and total tokens from the same runs, not a launch-page price.
More from Infra
- PostgreSQL 19 slips to November as LLM-assisted code review floods committers with edge cases — sanjaykalra · 2026-10-21
- Eon cuts engineer onboarding from 2 days to 2 minutes with an MCP Gateway on Bedrock AgentCore — AWS ML Blog · 2026-10-20
- a16z's Ben Horowitz says AI buildout isn't the 90s bubble — an investor pushes back — henloitsjoyce · 2026-10-20
- Micron and Nvidia drive ~36% of S&P 500's Q3 earnings growth — Beth_Kindig · 2026-10-20
- NVIDIA lands in Europe: Jensen Huang and German chancellor discuss industrial AI at GTC Berlin — nvidia · 2026-10-20
- Perplexity Computer now runs fully local on Mac, model and agent on-device — perplexity_ai · 2026-10-20