Gemini 3.7 Flash Review: Google's Coding & Agent Upgrade at Half Price — Pricing, Benchmarks & What's Next

Q: Gemini 3.7 Flash API Pricing

| Metric | Intro (until Dec 31, 2026) | Standard (from Jan 1, 2027) | |---|---|---| | Input | $0.75 / 1M tokens | $1.50 / 1M tokens | | Output | $3.75 / 1M tokens | $7.50 / 1M tokens | | Context window | 1M | 1M | > 💡 Pricing note: The introductory rate is exactly half of Gemini 3.6 Flash's launch pricing ($1.50 / $7.50). After December 31, 2026 it doubles — plan long-running production workloads with the 2027 rate in mind. At $0.75/$3.75, Gemini 3.7 Flash undercuts most mid-range rivals — but

Q: Gemini 3.7 Flash vs The Mid-Range Lineup

| Model | Input $ | Output $ | Context | Best for | |---|---|---|---|---| | Gemini 3.7 Flash (intro) | $0.75 | $3.75 | 1M | Coding, agents, high volume (until 12/31) | | Gemini 3.6 Flash | $1.50 | $7.50 | 1M | Same role, pre-price-cut benchmark | | GPT-5.6 Luna | $0.20 | $1.20 | 1.05M | Cheapest OpenAI, general chat | | GPT-5.6 Terra | $2.00 | $12.00 | 1.05M | Balanced intelligence/cost | | Claude Sonnet 5 | $3.00 | $15.00 | 1M | Speed + intelligence balance | | DeepSeek V4 Flash | $0.22 | $0.66

Q: What About Gemini 4? (And What Happened to 3.5 Pro?)

This is the bigger story behind the release: - Gemini 3.5 Pro is effectively cancelled. After repeated delays (missed its June launch, still in "partner testing" in July), reports on August 10 say Google DeepMind has dropped it entirely. - Gemini 4 officially replaces it — confirmed by Google on August 7. CEO Sundar Pichai said Google needs a *larger* Gemini 4 to compete at the frontier, calling coding and agentic systems the priority areas. - Cadence shift. Google CEO has talked up monthly rele

Q: How to Get Started with Gemini 3.7 Flash

1. API: model ID gemini-3.7-flash via ai.google.dev] — available now 2. AI Studio: try it free in Google AI Studio] 3. Frontier seat: Gemini API, Vertex AI, and Gemini Code Assist all support the new model 4. Agents: combine with Gemini Spark / agent runtime where available in your region

⚡ Quick Answer: Google released Gemini 3.7 Flash on August 13, 2026 — just three weeks after Gemini 3.6 Flash, its fastest release cadence yet. It's a meaningful coding and agentic-workflow upgrade: DeepSWE v1.1 climbs from 49.0% to 65.3%, it leads Zapier's AutomationBench, and the context window stays at 1M tokens. The hook is pricing: an introductory rate of $0.75 input / $3.75 output per 1M tokens — half of 3.6 Flash — but that discount expires December 31, 2026, after which rates double to $1.50 / $7.50.

What's New in Gemini 3.7 Flash

Google shipped Gemini 3.7 Flash on August 13, a mere three weeks after 3.6 Flash (July 21). For context, 3.6 Flash itself arrived just three weeks after the I/O-era 3.5 Flash lineup. Google has clearly moved Flash-tier releases to a monthly cadence — and it's using price cuts to win over developers mid-cycle.

What actually changed:

Coding. DeepSWE v1.1 score jumps from 49.0% to 65.3% — the biggest single jump in the Flash line.
Agents. Tops the public leaderboard on Zapier's AutomationBench (30.4%), which tests end-to-end workflows across 47 tools.
Documents & workflow automation. Substantial gains in document analysis and multi-step automation tasks.
1M token context — unchanged, but with unusually fast output generation.
Half-price intro. $0.75 / $3.75 per 1M tokens, exactly half of 3.6 Flash's launch price. No similar cut for the older model — this is an acquisition play.
Gemini Spark. Tied to Spark agent availability, though rollout is limited in some regions at launch.

Verdict from early reviews: a strong workhorse release — not a frontier leap, but a fast, cheap, genuinely better coder/agent model. The caveats: mixed results on non-coding benchmarks, and the half-price window has an expiry date.

Gemini 3.7 Flash API Pricing

Metric	Intro (until Dec 31, 2026)	Standard (from Jan 1, 2027)
Input	$0.75 / 1M tokens	$1.50 / 1M tokens
Output	$3.75 / 1M tokens	$7.50 / 1M tokens
Context window	1M	1M

💡 Pricing note: The introductory rate is exactly half of Gemini 3.6 Flash's launch pricing ($1.50 / $7.50). After December 31, 2026 it doubles — plan long-running production workloads with the 2027 rate in mind.

At $0.75/$3.75, Gemini 3.7 Flash undercuts most mid-range rivals — but note the competition below it: GPT-5.6 Luna ($0.20/$1.20) and DeepSeek V4 Flash ($0.22/$0.66, off-peak) are both cheaper still for high-volume workloads.

Gemini 3.7 Flash vs The Mid-Range Lineup

Model	Input $	Output $	Context	Best for
Gemini 3.7 Flash (intro)	$0.75	$3.75	1M	Coding, agents, high volume (until 12/31)
Gemini 3.6 Flash	$1.50	$7.50	1M	Same role, pre-price-cut benchmark
GPT-5.6 Luna	$0.20	$1.20	1.05M	Cheapest OpenAI, general chat
GPT-5.6 Terra	$2.00	$12.00	1.05M	Balanced intelligence/cost
Claude Sonnet 5	$3.00	$15.00	1M	Speed + intelligence balance
DeepSeek V4 Flash	$0.22	$0.66	1M	Budget workloads (off-peak)

Verdict: During the intro window, Gemini 3.7 Flash is arguably the best coding/agent value in the mid-range — DeepSWE 65.3% at $0.75 input is hard to beat. From January 2027, the math gets closer to 3.6 Flash's original positioning, and pure-volume users will still prefer Luna or DeepSeek.

What About Gemini 4? (And What Happened to 3.5 Pro?)

This is the bigger story behind the release:

Gemini 3.5 Pro is effectively cancelled. After repeated delays (missed its June launch, still in "partner testing" in July), reports on August 10 say Google DeepMind has dropped it entirely.
Gemini 4 officially replaces it — confirmed by Google on August 7. CEO Sundar Pichai said Google needs a larger Gemini 4 to compete at the frontier, calling coding and agentic systems the priority areas.
Cadence shift. Google CEO has talked up monthly releases — but analysts warn faster releases mean more testing and governance cost for enterprise buyers.

So 3.7 Flash is Google's current workhorse; Gemini 4 (expected Nov/Dec 2026 window per analyst inference) is the flagship everyone is waiting for. Track both on our free AI Model Release Tracker.

How to Get Started with Gemini 3.7 Flash

API: model ID gemini-3.7-flash via ai.google.dev — available now
AI Studio: try it free in Google AI Studio
Frontier seat: Gemini API, Vertex AI, and Gemini Code Assist all support the new model
Agents: combine with Gemini Spark / agent runtime where available in your region

Should You Switch?

From Gemini 3.6 Flash: Yes — same context, strictly better coding/agent scores, and half the price for the rest of 2026. No reason to stay unless you've pinned workflows.

From GPT-5.6 Luna / DeepSeek V4 Flash: Only if your workload is coding/agent-heavy. For pure volume or general chat, Luna ($0.20) and DeepSeek ($0.22) remain cheaper — and Gemini's intro price expires in December.

Waiting for a flagship? If you need frontier reasoning, Gemini 4 (or today's GPT-5.6 Sol / Claude Opus 5) beats anything in the Flash line. 3.7 Flash is the smart pick for high-volume coding pipelines, not for the hardest reasoning tasks.

FAQ

When was Gemini 3.7 Flash released?

August 13, 2026 — just three weeks after Gemini 3.6 Flash (July 21), Google's fastest-ever Flash-to-Flash release cycle.

How much does Gemini 3.7 Flash cost?

Introductory pricing is $0.75 input / $3.75 output per 1M tokens until December 31, 2026. From January 1, 2027 it doubles to $1.50 / $7.50.

Is Gemini 3.7 Flash better than Gemini 3.6 Flash?

Yes, especially for coding and agent workflows — DeepSWE v1.1 jumps from 49.0% to 65.3%, and it leads Zapier's AutomationBench. Both have 1M context.

What is Gemini 3.7 Flash's context window?

1M tokens — same as 3.6 Flash.

What happened to Gemini 3.5 Pro?

Cancelled per August 10 reports — after repeated delays, Google DeepMind reportedly dropped it. Google confirmed on August 7 that Gemini 4 will take its place as the next flagship.

When will Gemini 4 launch?

Not officially dated. Analyst estimates point to late 2026 (Nov/Dec window); Pichai has said Google needs a larger Gemini 4 to compete at the frontier.

Updated August 18, 2026 with data from Google announcements and launch-day reporting. Compare live prices on our AI Model Pricing page, or track every release on the AI Model Release Tracker. Related: Grok 4.6 Review, DeepSeek V4-Pro 0813 price increase, GPT-5.6 launch coverage.