Key Terms
- Gemini 3.7 Flash - "most intelligent workhorse model yet for coding and agents" (August 13), 1M context, 64K max output, tunable thinking (low/medium/high). FrontierCode 1.1 Main 43.6% (up from 3.6 Flash's 34.4%), Code Arena 1588 Elo. Sources: Google – Pricing , Local-Ai-Zone – Ai Updates August 2026
- Introductory pricing - 3.7 Flash at $0.75/$3.75 per MTok through December 31, 2026 (standard $1.50/$7.50 from January 1, 2027); the same intro price was extended retroactively to 3.6 Flash. Cached input $0.075. Non-global endpoints +10%. Source: Google – Pricing
- CodeMender - a Google code-repair product bundled into the same introductory-pricing banner as the Flash models ("Gemini 3.7 Flash, Gemini 3.6 Flash, and CodeMender using these models"). Source: Google – Pricing
- Priority / Flex / Batch tiers - Vertex service tiers: Priority at 1.8x standard (3.7 Flash intro: $1.35/$6.75), Flex/Batch at 50% (3.7 Flash intro: $0.375/$1.875). Source: Google – Pricing
Latest Changes
Changes since the 2026-07 report.
July watch items verified:
- Gemini 3.5 Pro GA (slipped twice in July): still-pending. The Vertex pricing page still lists Gemini 3.1 Pro Preview as the only Pro-tier option; no 3.5 Pro row exists. Source: Google – Pricing
- Gemini 3.6 Flash deprecation in Copilot (July 31): N/A. Google-side, 3.6 Flash remains available and now shares 3.7's intro price.
New August changes:
- New model (Aug 13): Gemini 3.7 Flash, three weeks after 3.6 Flash went stable. First-pass code accuracy up ~9 points on FrontierCode 1.1 Main; positioned for agentic coding where "quality matters but budgets are tight." Sources: Google – Pricing , Local-Ai-Zone – Ai Updates August 2026
- Price change: intro $0.75/$3.75 through Dec 31, 2026 for 3.7 Flash (half the standard rate) plus the retroactive extension of the same intro pricing to 3.6 Flash. Source: Google – Pricing
- Distribution: Gemini 3.7 Flash added to GitHub Copilot on August 13; Gemini 3.1 Pro was deprecated in Copilot on September 1 (Google-side it remains on Vertex as Preview). Sources: GitHub – 2026 08 13 Gemini 3 7 Flash Is Now Available In Github Copilot , GitHub – 2026 08 31 Selected Github Copilot Models Deprecated
Plans
Google AI consumer plans (free tier, Google AI Pro, Ultra) unchanged with no coding-agent limit disclosures. Gemini Code Assist pricing tiers (Free/Standard/Enterprise) unchanged; CodeMender surfaced in Vertex pricing materials. Sources: Google – Pricing , Gemini
API Pricing
| Model | Input $/MTok | Cached | Output $/MTok | Notes |
|---|---|---|---|---|
| Gemini 3.7 Flash (intro thru 2026-12-31) | $0.75 | $0.075 | $3.75 | Standard from 2027-01-01: $1.50/$7.50 |
| Gemini 3.7 Flash Priority | $1.35 | $0.135 | $6.75 | |
| Gemini 3.7 Flash Flex/Batch | $0.375 | $0.0375 | $1.875 | |
| Gemini 3.6 Flash | $0.75 / $3.75 (intro, same as 3.7) | $0.075 | Standard $1.50/$7.50 from 2027 | |
| Gemini 3.1 Pro Preview | $2.00 (<=200K) / $4.00 | $0.20/$0.40 | $12.00/$18.00 | Only Pro-tier option |
| Gemini 3.5 Flash | $1.50 | $0.15 | $9.00 | |
| Gemini 3.5 Flash-Lite | $0.30 | $0.03 | $2.50 | |
| Gemini 3.1 Flash-Lite | $0.25 | $0.025 | $1.50 |
Long-context: prompts over 200K billed at the higher rate for all tokens. Grounding: 5,000 free queries/month, then $14 per 1,000. Non-global endpoints +10% (GA from July 1, 2026). Source: Google – Pricing
Model Performance / Benchmarks
| Benchmark | Gemini 3.7 Flash | Gemini 3.6 Flash | Notes |
|---|---|---|---|
| FrontierCode 1.1 Main | 43.6% | 34.4% | Source: Local-Ai-Zone – Ai Updates August 2026 |
| Code Arena | 1588 Elo | - | Source: Local-Ai-Zone – Ai Updates August 2026 |
Latest News
Three-Week Iteration Cycle (August 13, 2026)
Gemini 3.7 Flash shipped 21 days after 3.6 Flash stabilized, cutting the effective price of the Flash tier in half (intro pricing) while raising FrontierCode accuracy ~9 points. The retroactive extension of intro pricing to 3.6 Flash blunts version-fragmentation cost concerns. Sources: Google – Pricing , Local-Ai-Zone – Ai Updates August 2026
Pro Tier Still Missing
Gemini 3.5 Pro, which slipped twice in July with no firm date, remained absent in August; Vertex's Pro option is still the 3.1 Pro Preview from February. Source: Google – Pricing
Community Signals
Aggregator commentary flagged the release-velocity problem: "Version fragmentation accelerating: 3-week gap between Gemini 3.6 and 3.7 Flash creates integration complexity." Source: Local-Ai-Zone – Ai Updates August 2026 . The Gemini CLI placed in the top five agent harnesses in August rankings. Source: Aitoolsrecap – Best Ai Coding Agent Harness 2026
Enterprise Readiness
| Feature | Available? | Details |
|---|---|---|
| SSO (SAML/OIDC) | Yes | Google Cloud IAM/Workspace identity. |
| SCIM | Yes | Cloud Identity/Workspace provisioning. |
| Audit logs | Yes | Cloud Audit Logs. |
| IP indemnity | Yes | Google's Gemini for Google Cloud indemnification program (carried over). Source: Google – Generative Ai Gdpr |
| Data residency | Yes | Global and regional endpoints with published 10% non-global premium. Source: Google – Pricing |
| HIPAA | Yes | Covered under Google Cloud HIPAA BAA. |
| Air-gapped / on-prem | Partial | Vertex AI on GDC (air-gapped) for select models. |
| SLA | Yes | Vertex AI SLAs. |
| Admin controls (RBAC) | Yes | IAM, CMEK, VPC-SC, model governance. |
Transparency Gaps
| Gap | Details | Severity |
|---|---|---|
| 3.5 Pro status | Two months of slipped dates with no public timeline; Pro-tier buyers are stuck on a February preview. | High |
| Intro-pricing cliff | Budgets built on $0.75/$3.75 double on January 1, 2027 for both Flash models. | Medium |
| Consumer coding limits | No Gemini Code Assist / consumer plan usage numbers. | Medium |
| CodeMender scope | Referenced in pricing but without a dedicated pricing or feature page. | Low |