Updated October 14, 2026 with Google's current Gemini API lineup and standard paid-tier prices, including Gemini 4 Argon's October 13 release and the October 8 deprecation of 3.7 Flash and 3.5 Flash. Prices are per 1 million tokens, for prompts up to 200k tokens, from Google's official pricing page.
Which Gemini model should I use right now?
Quick answer (October 2026)
- Default pick: Gemini 3.8 Flash. Google's most intelligent Flash model, at $0.75 input / $3.75 output, the same price as 3.6 Flash until December 31, 2026.
- Cheapest and fastest to respond: Gemini 3.1 Flash-Lite at $0.25 / $1.50, about 0.7s to first response. Use it for high-volume classification, extraction and translation.
- Smartest: Gemini 4 Argon (
gemini-4-argon), on the paid API tier since October 13, 2026, at $2 / $10 until November 30, then $4 / $20. It is not on the free tier. - Gemini 3.1 Pro Preview at $2 / $12 is still a preview model. It costs more than Argon until November 30 and less after.
How do current Gemini models compare on price and status?
| Model | Model ID | Status | Input per 1M | Output per 1M | Best for |
|---|---|---|---|---|---|
| Gemini 3.8 Flash | gemini-3.8-flash | Stable, newest Flash | $0.75 (from Jan 1, 2027: $1.50) | $3.75 (from Jan 1, 2027: $7.50) | Default for coding, agents and most app work |
| Gemini 3.7 Flash | gemini-3.7-flash | Deprecated Oct 8, 2026; requests auto-route to 3.8 Flash | n/a | n/a | Change the model ID to gemini-3.8-flash |
| Gemini 3.6 Flash | gemini-3.6-flash | Stable, previous generation | $0.75 (from Jan 1, 2027: $1.50) | $3.75 (from Jan 1, 2027: $7.50) | Replacement for 3.5 Flash; no price reason to pick it over 3.8 Flash |
| Gemini 3.5 Flash | gemini-3.5-flash | Deprecated Oct 8, 2026; requests auto-route to 3.6 Flash | n/a | n/a | Change the model ID to gemini-3.6-flash or gemini-3.8-flash |
| Gemini 3.5 Flash-Lite | gemini-3.5-flash-lite | Stable | $0.30 | $2.50 | Fast, cheap agent steps and bulk processing |
| Gemini 3.1 Flash-Lite | gemini-3.1-flash-lite | Stable; shutdown earliest May 7, 2027 (Google recommends 3.5 Flash-Lite) | $0.25 ($0.50 audio) | $1.50 | Lowest cost per token |
| Gemini 3 Flash Preview | gemini-3-flash-preview | Preview, legacy | $0.50 ($1 audio) | $3.00 | Older preview; prefer a stable Flash |
| Gemini 3.1 Pro Preview | gemini-3.1-pro-preview | Preview | $2 (over 200k prompt: $4) | $12 (over 200k prompt: $18) | Hardest reasoning where quality beats cost |
| Gemini 4 Argon | gemini-4-argon | Paid tier since Oct 13, 2026; not on the free tier | $2 (from Dec 1, 2026: $4) | $10 (from Dec 1, 2026: $20) | Hardest coding, reasoning and long agent tasks |
What this means in practice: 3.6 and 3.8 Flash cost the same, and calls to 3.7 Flash and 3.5 Flash are already served by 3.8 Flash and 3.6 Flash, so start new work on 3.8 Flash. Drop to a Flash-Lite model only when cost or volume is the bottleneck, and move to 3.1 Pro only for tasks where Flash answers are not good enough. Batch and Flex pricing halve most of these rates.
Which Gemini model is fastest?
OpenRouter publishes median (P50) latency and throughput from real API traffic. These are its Google AI Studio numbers, read on October 8, 2026. Latency is the time until the response starts; throughput is output tokens per second while it streams. They mix all prompt types and reasoning settings, so treat them as a rough live comparison, not a controlled benchmark.
| Model | Latency (P50) | Throughput (P50) |
|---|---|---|
| Gemini 3.8 Flash | 2.61s | 78 tok/s |
| Gemini 3.7 Flash (now routes to 3.8 Flash) | 2.33s | 79 tok/s |
| Gemini 3.6 Flash | 1.04s | 82 tok/s |
| Gemini 3.5 Flash (now routes to 3.6 Flash) | 1.79s | 169 tok/s |
| Gemini 3.5 Flash-Lite | 0.74s | 114 tok/s |
| Gemini 3.1 Flash-Lite | 0.67s | 134 tok/s |
| Gemini 3 Flash Preview | 1.30s | 82 tok/s |
| Gemini 3.1 Pro Preview | 3.78s | 106 tok/s |
The Flash-Lite models start responding fastest, in under a second. 3.7 and 3.8 Flash take over 2 seconds to start, probably because more of their requests use reasoning. 3.5 Flash streamed fastest once it started, but its requests now go to 3.6 Flash. For our controlled copywriting test of 3.5 Flash, 3 Flash and 3.1 Flash-Lite, see the Gemini 3.5 Flash speed and cost benchmark.
Should I pick Flash, Flash-Lite or Pro?
This is a tier choice, not a version-number choice. A newer Flash-Lite is not a better Flash; it is a cheaper tier with a different job.
- Pro (3.1 Pro Preview) is quality-first: long reasoning, hard coding and analysis where a wrong answer costs more than tokens.
- Flash (3.8 Flash) is the balanced default: strong enough for most coding, agent and writing work at a fraction of Pro's price.
- Flash-Lite (3.5 or 3.1 Flash-Lite) is cost-first: high-volume, simple, repeatable tasks.
Earlier versions of this comparison (Gemini 3.1 Pro vs 3.1 Flash-Lite vs 3 Flash, and Flash vs Pro) are merged into this page. Their models all appear in the table above with current prices.
Which older Gemini models are retired or restricted?
gemini-3.7-flashandgemini-3.5-flashwere deprecated on October 8, 2026. Requests are automatically routed togemini-3.8-flashandgemini-3.6-flash. Update your model ID.gemini-3.1-flash-litecan shut down as early as May 7, 2027. Google recommendsgemini-3.5-flash-lite.gemini-2.0-flashshut down on June 1, 2026.gemini-3.1-flash-lite-previewshut down on May 25, 2026. Usegemini-3.1-flash-lite.- Gemini 2.5 Flash, 2.5 Flash-Lite and 2.5 Pro are not deprecated, but Google limits them to accounts that used them before. New projects should use 3.5 Flash-Lite or 3.8 Flash.
What did the original 2025 ranking show?
This page started as a September 2025 Gemini 2.0/2.5 ranking. We could not trace its original benchmark method, so treat these figures as unverified history, not a current recommendation.
| Model | Latency (s) | Throughput (tps) | Max Output Tokens | Input Cost per 1M | Output Cost per 1M |
|---|---|---|---|---|---|
| Gemini 2.5 Flash Lite | 0.59 | 178.70 | 66,000 | $0.10 | $0.40 |
| Gemini 2.0 Flash (retired) | 0.53 | 153.10 | 8,000 | $0.15 | $0.60 |
| Gemini 2.5 Flash | 0.65 | 97.39 | 66,000 | $0.30 | $2.50 |
| Gemini 2.5 Pro | 2.22 | 85.37 | 66,000 | $1.25 to $2.50 | $10 to $15 |
Frequently Asked Questions
What is the newest Gemini model?
Gemini 4 Argon (gemini-4-argon), available on the paid Gemini API tier since October 13, 2026. The newest Flash model is Gemini 3.8 Flash (gemini-3.8-flash).
Is Gemini 3.8 Flash better than Gemini 3.6 Flash?
Google describes 3.8 Flash as its most intelligent Flash model. Both cost $0.75 input and $3.75 output per 1M tokens until December 31, 2026, so there is no price reason to stay on 3.6 Flash. Requests to the deprecated 3.5 Flash now go to 3.6 Flash automatically.
When is Gemini 4 available in the API?
Since October 13, 2026, on the paid tier only, as gemini-4-argon. It costs $2 input and $10 output per 1M tokens through November 30, 2026, then $4 and $20.
Which Gemini model is cheapest?
Gemini 3.1 Flash-Lite, at $0.25 input and $1.50 output per 1M text tokens on the standard paid tier. Batch requests cost half.
How do I calculate my monthly Gemini cost?
Multiply input tokens by the input price per 1M and output tokens by the output price per 1M, then add them. For example, 10M input and 2M output tokens on 3.8 Flash cost 10 × $0.75 + 2 × $3.75 = $15 before December 31, 2026.
Sources
- https://ai.google.dev/gemini-api/docs/models
- https://ai.google.dev/gemini-api/docs/pricing
- https://ai.google.dev/gemini-api/docs/deprecations
- https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/
- https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-argon/
- https://aistudio.google.com/docs/pricing
- OpenRouter model pages (live latency and throughput by provider)