Gemini 3.7 Flash Review 2026: Deprecated AI Coding Model Guide
Gemini 3.7 Flash was Google DeepMind's coding-focused model -- pricing, specs, and why it is now deprecated in favor of Gemini 3.8 Flash.
Categories & Tags
Gemini 3.7 Flash: Google DeepMind's Short-Lived Coding Model
Gemini 3.7 Flash is a model Google DeepMind published on August 13, 2026 as part of the Gemini 3 family, built directly on Gemini 3.6 Flash. It is no longer the model to reach for: Google's own Gemini API changelog confirms it was deprecated on October 8, 2026, replaced by Gemini 3.8 Flash, pulled from selection in Google AI Studio, and that any API request still using the gemini-3.7-flash model ID is now automatically routed to gemini-3.8-flash. This review covers what it offered during its roughly two-month run, its launch pricing, and why developers today should target Gemini 3.8 Flash directly instead.
Quick Navigation
Quick Summary
Overall Rating 4.0/5 Best For Historical reference only — it was a strong value coding/agentic model at launch, now deprecated. Pricing (at launch) $0.75 per 1M input tokens / $3.75 per 1M output tokens (introductory rate). Free Plan Yes, rate-limited, in Google AI Studio and the Gemini API. Current Status Deprecated since October 8, 2026 — replaced by Gemini 3.8 Flash. Ease of Use 4.2/5 (API-based access via Google AI Studio, no separate install).
What Is Gemini 3.7 Flash?
Gemini 3.7 Flash is a language model Google DeepMind published on August 13, 2026 as part of the Gemini 3 family. It builds directly on Gemini 3.6 Flash, sharing the same underlying architecture and training data. It has since been deprecated: per Google's Gemini API changelog entry dated October 8, 2026, Gemini 3.7 Flash has been replaced by Gemini 3.8 Flash, is no longer selectable in Google AI Studio, and API requests using the gemini-3.7-flash model ID are now automatically routed to gemini-3.8-flash. For developers exploring other coding-focused assistants, theaitoolsbox.com also covers GitHub Copilot and Cursor in our AI Coding Tools category.
Who Gemini 3.7 Flash Was For
When it was live, Gemini 3.7 Flash was aimed at developers and teams who wanted strong measured intelligence at a low per-token cost, with the flexibility to tune quality against cost and latency through customizable thinking configurations. Its 1 million token input window and 64,000 token output ceiling suited workloads involving large volumes of text, images, audio, video, or documents. It was reachable through the Gemini API, Google AI Studio, the Gemini app via the Spark agent, Gemini Enterprise App, Gemini Enterprise Agent Platform, and Google Antigravity, so it fit both API-first builders and teams working inside Google's hosted surfaces.
Reality check: None of that matters for a new build today — the model ID is deprecated and auto-routes to Gemini 3.8 Flash, so there's no reason to target "gemini-3.7-flash" specifically in new code.
Gemini 3.7 Flash Features
Reasoning upgrade over 3.6 Flash
Relative to Gemini 3.6 Flash, Gemini 3.7 Flash introduced algorithmic improvements to its core reasoning foundation and added support for agentic video understanding.
Built on the same base model, with a focused reasoning upgrade
Customizable thinking configurations
It also supports customizable thinking configurations, a hybrid reasoning setup that lets developers tune the tradeoff between quality, cost, and latency per request.
Dial response quality up or down against cost and speed
1M token input, 64K output
The model accepted up to 1 million tokens of input across text, images, audio, video, and documents, and could produce up to 64,000 output tokens.
Enough context for large codebases or long documents
Available across Google's surfaces
Distribution covered the Gemini app through the Spark agent, the Gemini Enterprise App, the Gemini Enterprise Agent Platform, Google AI Studio, the Gemini API, and Google Antigravity, Google's agentic coding IDE.
Reachable from both API-first and hosted workflows
Gemini 3.7 Flash Pricing (Historical)
On the paid Gemini API tier, Gemini 3.7 Flash launched at an introductory $0.75 per 1 million input tokens and $3.75 per 1 million output tokens — a 50% cut against Gemini 3.6 Flash's standard rate. That introductory pricing was scheduled to run through December 31, 2026, with the standard rate set to rise to $1.50 per 1M input tokens and $7.50 per 1M output tokens from January 1, 2027. Cached input tokens cost $0.075 per 1M tokens (a 90% discount), and a free tier with rate limits was available in Google AI Studio and the Gemini API. None of this applies to new usage anymore: because the model is deprecated as of October 8, 2026, any request using the gemini-3.7-flash model ID is automatically routed to gemini-3.8-flash, which bills at Gemini 3.8 Flash's own current rate.
| Plan (as launched) | Price | What You Got |
|---|---|---|
| Free Tier | $0 | Rate-limited access in Google AI Studio and the Gemini API. |
| Pay-as-you-go (introductory) Launch Rate | $0.75 / $3.75 per 1M tokens | Input / output rate through Dec 31, 2026; cached input $0.075 per 1M tokens. |
| Pay-as-you-go (standard, from Jan 1, 2027) | $1.50 / $7.50 per 1M tokens | Scheduled standard rate — moot now since the model ID is deprecated. |
For current, live pricing, use the official Gemini API pricing page, which reflects Gemini 3.8 Flash's own rate.
Where Gemini 3.7 Flash Was Strong / Where It Fell Short
- Reasoning gain over 3.6 FlashScored 56 on the Artificial Analysis Intelligence Index in Google DeepMind's own published table, against 52 for Gemini 3.6 Flash.
- Aggressive launch pricingAt $0.75/$3.75 per 1M tokens it undercut Claude Sonnet 5 ($2.00/$10.00) and GPT-5.6 Terra ($2.00/$12.00) by a wide margin while scoring close to Claude Sonnet 5's 55.
- Large context window1 million input tokens and 64,000 output tokens across text, image, audio, and video input.
- Deprecated after ~2 monthsLaunched Aug 13, 2026, deprecated Oct 8, 2026 — one of the shortest lifespans of any Gemini Flash release.
- Not a leading benchmark scoreIts 56 on the Artificial Analysis Intelligence Index landed roughly level with, not clearly ahead of, Claude Sonnet 5's 55.
- No longer selectablePulled from Google AI Studio's model picker entirely after deprecation.
Deprecation Status: What Changed on October 8, 2026
Google's Gemini API changelog entry dated October 8, 2026 confirms Gemini 3.7 Flash is deprecated and has been replaced by Gemini 3.8 Flash. It is no longer available to select in Google AI Studio, and any API request that still specifies the gemini-3.7-flash model ID is automatically routed to gemini-3.8-flash instead. Teams with that model string hardcoded anywhere in their code should update it directly to avoid relying on the auto-routing behavior, and re-run their evaluations against the new model's outputs before shipping.
Is Gemini 3.7 Flash Worth Targeting in 2026?
No, not as a specified model. Gemini 3.7 Flash was a strong value option at launch, pairing a 56 on the Artificial Analysis Intelligence Index with introductory pricing well below Claude Sonnet 5 and GPT-5.6 Terra. But since it is deprecated and requests to the gemini-3.7-flash model ID now route automatically to gemini-3.8-flash, there is no practical reason to target the old model ID in new code. A developer starting today should target Gemini 3.8 Flash directly, or compare it against other AI coding tools before committing.
Gemini 3.7 Flash vs the Competition (At Launch)
| Decision Area | Gemini 3.7 Flash | When Another Option Wins |
|---|---|---|
| Measured intelligence | 56 on the Artificial Analysis Intelligence Index, ahead of Gemini 3.6 Flash's 52. | GPT-5.6 Terra scored higher on the same published index, if raw intelligence mattered more than price. |
| Price per token | $0.75/$3.75 per 1M tokens (introductory) — far below Claude Sonnet 5 ($2.00/$10.00) and GPT-5.6 Terra ($2.00/$12.00). | None of the three competitors beat Gemini 3.7 Flash on launch price, per the published table. |
| Availability today | Deprecated — the model ID auto-routes to Gemini 3.8 Flash as of Oct 8, 2026. | Every actively maintained competitor (Claude Sonnet 5, GPT-5.6 Terra, Gemini 3.8 Flash itself) wins here simply by still being the current model. |
Gemini 3.7 Flash vs Gemini 3.8 Flash
Gemini 3.8 Flash is the direct successor and the model every Gemini 3.7 Flash API call now actually runs on, since deprecated requests auto-route to it. There is no scenario where continuing to target the old model ID beats simply calling Gemini 3.8 Flash directly — you get the same routing destination with none of the deprecation risk of a model string that could stop being honored in a future API version.
Choose Gemini 3.7 Flash if: You're researching historical benchmark or pricing context for a past project. Choose Gemini 3.8 Flash if: You're building anything new — see our Gemini 3.8 Flash review for current specs and pricing.
Frequently Asked Questions
What is Gemini 3.7 Flash?
Gemini 3.7 Flash is a Google DeepMind language model published August 13, 2026, part of the Gemini 3 family and built on Gemini 3.6 Flash, offering a 1 million token context window and customizable thinking configurations.
How much did Gemini 3.7 Flash cost?
Yes, it had published pricing: $0.75 per 1 million input tokens and $3.75 per 1 million output tokens on the introductory rate, scheduled to rise to $1.50/$7.50 per 1M tokens on January 1, 2027 before the model was deprecated.
Is Gemini 3.7 Flash still available?
No. Google deprecated Gemini 3.7 Flash on October 8, 2026. It is no longer selectable in Google AI Studio, and API requests to the gemini-3.7-flash model ID are automatically routed to gemini-3.8-flash.
What replaced Gemini 3.7 Flash?
Gemini 3.8 Flash replaced it. Developers who had gemini-3.7-flash hardcoded in their applications should update the model string directly rather than rely on Google's automatic routing.
Key Takeaways
- Gemini 3.7 Flash launched August 13, 2026 at an introductory $0.75/$3.75 per 1M tokens, cheaper than Claude Sonnet 5 and GPT-5.6 Terra.
- It scored 56 on the Artificial Analysis Intelligence Index with a 1M-token context window and 64K-token output.
- It was deprecated October 8, 2026 — roughly two months after launch — and is now auto-routed to Gemini 3.8 Flash.
Gemini 3.7 Flash Alternatives
- Gemini 3.8 Flash — The actual current model: every deprecated 3.7 Flash request already auto-routes here, and it keeps the same 1M-token context window while billing at its own current per-token rate rather than 3.7 Flash's old $0.75/$3.75 pricing.
- GitHub Copilot — An IDE-native coding assistant billed at a flat $10/month per-seat subscription rather than Gemini 3.7 Flash's per-token API pricing, so costs don't scale with usage volume.
- Cursor — An AI-native code editor built around multi-model access, letting you switch between models like Claude Sonnet 5 or GPT-5.6 Terra in one IDE instead of being locked to a single vendor's deprecation schedule.
- Claude — Anthropic's Claude Sonnet 5 scored 55 on the same Artificial Analysis Intelligence Index Gemini 3.7 Flash scored 56 on, at a higher $2.00/$10.00 per 1M token rate.
For broader context, see our roundups of the best AI coding tools in 2026, our head-to-head on Google Gemini vs ChatGPT, and our frontier-model comparison of GPT-6 Astra vs Claude Fable 5.1.
Bottom Line: Gemini 3.7 Flash was a short-lived, competitively priced model from Google DeepMind that improved on Gemini 3.6 Flash's reasoning, added agentic video understanding and customizable thinking configurations, and offered a 1M-token input window with 64K output at an introductory $0.75/$3.75 per 1M tokens. It scored 56 on the Artificial Analysis Intelligence Index, roughly level with Claude Sonnet 5 while costing far less. But it was deprecated on October 8, 2026 after roughly two months, replaced by Gemini 3.8 Flash, with its model ID now auto-routing to the successor. Treat it as historical context, not a model to build on.
Last Reviewed: October 2026 | Reviewed by theaitoolsbox.com editorial team
Quick Answers
What is Gemini 3.7 Flash?
Gemini 3.7 Flash is a Google DeepMind language model published August 13, 2026, part of the Gemini 3 family and built on Gemini 3.6 Flash, with a 1 million token context window.
How much did Gemini 3.7 Flash cost?
Yes, it had published pricing of $0.75 per 1 million input tokens and $3.75 per 1 million output tokens on its introductory rate, through December 31, 2026.
Is Gemini 3.7 Flash still available?
No, Google deprecated Gemini 3.7 Flash on October 8, 2026, and API requests to the gemini-3.7-flash model ID now automatically route to Gemini 3.8 Flash.
Gemini 3.7 Flash
AI Coding Tools
Pricing Plans
1st Free Subscription
Various plans available
Free Tier
$0Rate-limited access in Google AI Studio and the Gemini API.
Pay-as-you-go (introductory)
$0.75 / $3.75 per 1M tokensInput / output rate through Dec 31, 2026; cached input $0.075 per 1M tokens.
Pay-as-you-go (standard, from Jan 1, 2027)
$1.50 / $7.50 per 1M tokensScheduled standard rate -- moot now since the model ID is deprecated.