Best LLMs for Instruction Following — Google
The best Google model for instruction following is Gemini 3.1 Flash Lite, satisfying 77.2% of IFBench's verifiable prompt constraints. Gemini 3.1 Flash Lite Preview (77.2%) and Gemini 3.1 Pro Preview (77.1%) round out the top three.
- 1
gemini-3.1-flash-liteReasoningToolsJSON+225.0 intel · $0.250/M · 337 t/s77.2%IFBench - 2gemini-3.1-flash-lite-previewReasoningToolsJSON+225.0 intel · $0.250/M · 337 t/s77.2%IFBench
- 3gemini-3.1-pro-previewReasoningToolsJSON+246.5 intel · $2.00/M · 136 t/s77.1%IFBench
- 4gemini-3.5-flashReasoningToolsJSON+250.2 intel · $1.50/M · 275 t/s76.3%IFBench
- 5gemini-2.5-proReasoningToolsJSON+225.8 intel · $1.25/M · 113 t/s48.7%IFBench
- 6gemini-2.5-flashReasoningToolsJSON+214.1 intel · $0.300/M · 75 t/s39.0%IFBench
- 7gemini-2.5-flash-liteReasoningToolsJSON+26.9 intel · $0.100/M · 91 t/s31.5%IFBench
How this is ranked
AI models ranked by IFBench — how reliably a model obeys explicit, verifiable constraints in the prompt (format, length, inclusion and exclusion rules). High scores mean fewer retries and less prompt-wrangling in production, which often matters more than raw intelligence for pipelines.
Frequently asked
Which Google model follows instructions most reliably?
The best Google model for instruction following is Gemini 3.1 Flash Lite, satisfying 77.2% of IFBench's verifiable prompt constraints. Gemini 3.1 Flash Lite Preview (77.2%) and Gemini 3.1 Pro Preview (77.1%) round out the top three.
What's a good alternative to Gemini 3.1 Flash Lite?
Gemini 3.1 Flash Lite Preview (77.2%) is the closest alternative on this metric, followed by Gemini 3.1 Pro Preview (77.1%). See the full ranking above for the tradeoffs.
How many Google models are there?
modelgrep tracks 30 Google models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by Gemini 3.5 Flash. 7 of them qualify for this ranking.