Key Terms
- Token-based billing - charges based on the number of input and output tokens processed. One token is roughly 4 characters. Gemini 3.x models use a 200K-token threshold: requests at or below 200K input tokens are billed at the standard rate, while larger requests are billed at a higher rate. Source: Google – Pricing
- Prompt caching - stores frequently used prompt prefixes to reduce cost. Cached input tokens cost roughly 90% less than standard input tokens. Available for all Gemini 3.x and 2.5 models. Source: Google – Pricing
- Context window - the maximum tokens a model can process in one conversation. Gemini 3.5 Flash, 3.1 Pro, and 3 Flash support a 1M token context window. Source: Google – Code Assist
- Antigravity CLI - Google's Go-based terminal AI agent that replaced Gemini CLI for free and individual users on June 18, 2026. It orchestrates multiple agents asynchronously and shares the same agent harness as the Antigravity 2.0 desktop app. Source: Googleblog – An Important Update Transitioning Gemini Cli To Antigravity Cli
- Antigravity 2.0 - Google's agent-first development platform, announced at I/O 2026, including a standalone desktop app for orchestrating cohorts of autonomous AI agents. Source: Google – Sundar Pichai Io 2026
- Interactions API - Google's now-GA unified endpoint for Gemini models and agents, with server-side state, background execution, Managed Agents, and a typed "Steps" schema replacing the old role structure. It is the default for Google AI Studio, the Gemini API, and all documentation. Source: Google – Interactions Api General Availability
- Managed Agents - an Interactions API feature (GA) that provisions a remote Linux sandbox where an agent can reason, run code, browse, and manage files; the Antigravity agent ships as the default and custom agents are defined as markdown files. Source: Google – Interactions Api General Availability
- Agent mode - a multi-step collaborative reasoning agent inside Gemini Code Assist chat that can edit multiple files, use full project context, call built-in tools, and integrate MCP servers. Available on both Standard and Enterprise. Source: Google – Code Assist
- Global vs Non-global pricing - Gemini 3.5 Flash and 3.1 Flash-Lite offer lower pricing when requests are served from Google's global compute pool, and 10% higher pricing when routed to a specific non-global region (e.g. Flash non-global input is $1.65/MTok vs $1.50 global). Source: Google – Pricing
- Priority tier - a premium serving option with higher availability and lower latency at 1.8x the standard price, available for all Gemini 3.x models. Source: Google – Pricing
- Flex/Batch tier - a discounted serving option for non-urgent workloads at roughly 50% of standard pricing (e.g. Gemini 3.5 Flash Flex/Batch input is $0.75/MTok vs $1.50 standard). Source: Google – Pricing
Latest Changes
Changes since the 2026-05 report.
- Verified (May watch-item): Free Gemini CLI and individual Code Assist shut down on June 18. The transition happened on schedule. Google's developer blog states Gemini CLI and the Gemini Code Assist IDE extensions "will stop serving requests" for Google AI Pro, Ultra, and free individual users on June 18, 2026; those users must migrate to Antigravity CLI. Enterprise Code Assist (Standard/Enterprise license) access is unchanged. Source: Googleblog – An Important Update Transitioning Gemini Cli To Antigravity Cli
- Verified (May watch-item): Gemini 3.5 Pro did NOT launch in June; it slipped to July. Business Insider reported on June 24 that Google pushed the Gemini 3.5 Pro launch to July, reversing the I/O promise that it would land "next month." Google is using the extra time to gather real-world feedback from early testers (the model has been spotted on Antigravity and LMArena) and to fold in criticism that Flash 3.5 consumed tokens too quickly. No price has been announced. Source: Businessinsider – Google 3 5 Pro July Release Tokens Ai Agents Model 2026 6
- Deprecation (consumer GitHub app): The consumer version of Gemini Code Assist on GitHub was deprecated June 18, 2026 (no new installations), with a hard full shutdown on July 17, 2026 when all code review activities end. The enterprise version is unaffected. Source: Google – Consumer Code Review
- Product GA: Interactions API reached general availability as Google's "primary API for interacting with Gemini models and agents," with a stable schema, Managed Agents (Antigravity agent as the default in a remote Linux sandbox), background execution, mixed built-in tool calling, Deep Research upgrades, media generation (Nano Banana 2 images, Lyria 3 music, multi-speaker TTS), Flex (50% reduction) and Priority tiers, and 55-day interaction retention on the paid tier. All documentation now defaults to it; the legacy generateContent API remains supported but frontier agentic capabilities will increasingly land only on Interactions API. Source: Google – Interactions Api General Availability
- Schema change: Interactions API moved "from Roles to Steps," replacing the old role structure with typed steps (user_input, thought, function_call, model_output, etc.). Google published a migration guide; a breaking-changes notice is dated May 2026. Source: Google – Interactions Api General Availability
- Product update: Gemini Spark reached macOS (June 30, beta) for Google AI Ultra subscribers aged 18 and over in the US, adding connected apps (Canva, Dropbox, Instacart, OpenTable, Zillow Rentals, Google Tasks, Google Keep), custom Model Context Protocol (MCP) support, real-time topic tracking, and a "coming soon" ability to assign multi-step tasks to a Mac remotely from a phone. Source: Google – Gemini Spark Updates June 2026
- Enforcement signal: Google has been banning Gemini/Antigravity Pro subscription accounts when used through non-approved third-party tools, an HN-submitted Reddit thread from February 2026 documents, with affected users reporting bans even when only using opencode (not third-party "clawd/molt" wrappers). This hardens the ecosystem boundary the June 18 cutover formalized. Source: News – Item
- Pricing verified unchanged: Gemini 3.5 Flash remains $1.50 input / $9.00 output per MTok (Global) on the live Vertex AI pricing page. Code Assist Standard stays $22.80/user/month ($19 annual) and Enterprise $54/user/month ($45 annual). Sources: Google – Pricing , Google – Pricing
- Partner model on Vertex (noted): The Vertex AI pricing page now lists Claude Sonnet 5 (with a note that standard pricing begins September 1, 2026) in the Model Garden, meaning Google Cloud customers can buy Anthropic models alongside Gemini on the same platform. This does not change Gemini Code Assist's own model lineup. Source: Google – Pricing
- SDK release: Google released ADK Go 2.0 (June 30), a graph-based workflow engine for multi-agent applications with built-in human-in-the-loop and dynamic orchestration. Source: Googleblog – Announcing Adk Go 20
Plans
| Plan | Price (monthly) | Price (annual) | Key Inclusions |
|---|---|---|---|
| Gemini Code Assist for individuals | $0 | N/A | Code completion, generation, chat in IDE; served through Antigravity CLI after June 18 (formerly Gemini CLI); local codebase awareness. Via Google Developer Program. Usage limits: undisclosed. Source: Google – Overview |
| Gemini Code Assist Standard | $22.80/user/month | $19/user/month | Code completion/generation/chat, local codebase awareness, code transformation, Gemini in Colab Enterprise, database dev assistance, Gemini in Firebase, BigQuery data insights, code completion in Cloud Run, enterprise-grade security, Gen AI indemnification, agent mode (preview), Gemini CLI. 30-day free trial for up to 50 users |
| Gemini Code Assist Enterprise | $54/user/month | $45/user/month | Everything in Standard + code customization (private repos), Gemini in Apigee, Gemini in Application Integration, Gemini Cloud Assist (preview, no additional charge), increased agent mode usage, increased daily usage limits |
Hourly list prices (how Google Cloud bills commitments): Standard $0.031232877/hour (monthly) and $0.026027397/hour (12-month); Enterprise $0.073972603/hour (monthly) and $0.061643836/hour (12-month). At 730 hours/month these resolve to the $22.80/$19 Standard and $54/$45 Enterprise figures above. Source: Google – Pricing
Terms explained:
- Code customization - tailors code suggestions to an organization's private codebases. Enterprise-only. Source: Google – Pricing
- Gemini Cloud Assist - AI-powered infrastructure design, diagnostics, cost optimization, and troubleshooting. Preview, at no additional charge for Enterprise customers. Source: Google – Pricing
- Gen AI indemnification - Google covers legal costs if Gemini Code Assist output infringes third-party copyright. Available on both Standard and Enterprise. Source: Google – Protecting Customers With Generative Ai Indemnification
Source: Google – Pricing , Google – Code Assist
API Pricing
Gemini 3.5 Flash pricing was re-verified on the live Vertex AI pricing page (Input $1.50/MTok Global, Output $9.00/MTok Global) and is unchanged from May. The full Gemini 3.x/2.5 lineup below is the current published state on that same live page.
Standard Pricing
| Model | Input ($/MTok) <= 200K | Input ($/MTok) > 200K | Cached Input ($/MTok) <= 200K | Cached Input ($/MTok) > 200K | Output ($/MTok) <= 200K | Output ($/MTok) > 200K |
|---|---|---|---|---|---|---|
| Gemini 3.5 Flash (Global) | $1.50 | $1.50 | $0.15 | $0.15 | $9.00 | $9.00 |
| Gemini 3.5 Flash (Non-global) | $1.65 | $1.65 | $0.165 | $0.165 | $9.90 | $9.90 |
| Gemini 3.1 Pro Preview | $2.00 | $4.00 | $0.20 | $0.40 | $12.00 | $18.00 |
| Gemini 3 Flash Preview (text/image/video) | $0.50 | $0.50 | $0.05 | $0.05 | $3.00 | $3.00 |
| Gemini 3 Flash Preview (audio) | $1.00 | $1.00 | $0.10 | $0.10 | $3.00 | $3.00 |
| Gemini 3.1 Flash-Lite (Global) | $0.25 | $0.25 | $0.025 | $0.025 | $1.50 | $1.50 |
| Gemini 3.1 Flash-Lite (Non-global) | $0.275 | $0.275 | $0.0275 | $0.0275 | $1.65 | $1.65 |
| Gemini 2.5 Pro | $1.25 | $2.50 | $0.13 | $0.25 | $10.00 | $15.00 |
| Gemini 2.5 Flash (text/image/video) | $0.30 | $0.30 | $0.03 | $0.03 | $2.50 | $2.50 |
Flex/Batch Pricing (~50% discount)
| Model | Input ($/MTok) | Output ($/MTok) |
|---|---|---|
| Gemini 3.5 Flash (Global) | $0.75 | $4.50 |
| Gemini 3.1 Pro Preview (<=200K) | $1.00 | $6.00 |
| Gemini 3.1 Pro Preview (>200K) | $2.00 | $9.00 |
| Gemini 3 Flash Preview (text) | $0.25 | $1.50 |
| Gemini 3.1 Flash-Lite (Global) | $0.125 | $0.75 |
Priority Pricing (1.8x standard)
| Model | Input ($/MTok) <= 200K | Input ($/MTok) > 200K | Output ($/MTok) <= 200K | Output ($/MTok) > 200K |
|---|---|---|---|---|
| Gemini 3.5 Flash (Global) | $2.70 | $2.70 | $16.20 | $16.20 |
| Gemini 3.1 Pro Preview | $3.60 | $7.20 | $21.60 | $32.40 |
| Gemini 3 Flash Preview | $0.90 | $0.90 | $5.40 | $5.40 |
| Gemini 3.1 Flash-Lite (Global) | $0.45 | $0.45 | $2.70 | $2.70 |
Notes:
- Gemini 3.5 Pro is not yet priced because the model is not GA. Pre-release access has been spotted in the Antigravity CLI and on LMArena, but Google has published no official API price. Source: Businessinsider – Google 3 5 Pro July Release Tokens Ai Agents Model 2026 6
- Vertex AI Model Garden now also sells partner models (e.g. Claude Sonnet 5), with standard pricing beginning September 1, 2026. Source: Google – Pricing
Source: Google – Pricing
Model Performance / Benchmarks
Google has published no new Gemini benchmark scores in June, because Gemini 3.5 Pro, the model that would carry fresh headline numbers, slipped to July. The most recent verifiable coding/agentic benchmark figures remain the Gemini 3.5 Flash numbers from the May I/O launch.
| Benchmark / Metric | Score | Notes |
|---|---|---|
| Terminal-Bench 2.1 (Gemini 3.5 Flash) | 76.2% | Published at I/O 2026. Source: Google – Gemini 3 5 |
| GDPval-AA (Gemini 3.5 Flash) | 1656 Elo | Published at I/O 2026. Source: Google – Gemini 3 5 |
| MCP Atlas (Gemini 3.5 Flash) | 83.6% | Published at I/O 2026. Source: Google – Gemini 3 5 |
| CharXiv Reasoning (Gemini 3.5 Flash) | 84.2% | Published at I/O 2026. Source: Google – Gemini 3 5 |
| Output speed (Gemini 3.5 Flash) | 4x faster than other frontier models | Google claim at I/O 2026. Source: Google – Gemini 3 5 |
| Antigravity-optimized 3.5 Flash speed | 12x faster than other frontier models | Within the Antigravity platform. Source: Google – Sundar Pichai Io 2026 |
| Gemini 3.5 Pro benchmarks | undisclosed (not yet GA) | No official scores published; model slipped to July. Source: Businessinsider – Google 3 5 Pro July Release Tokens Ai Agents Model 2026 6 |
Google showed comparison charts at I/O as animated GIFs rather than tabular data, so exact scores versus Claude Opus 4.8, Sonnet 5, GPT-5.5, and GLM-5.2 are not extractable from official sources. Independent community benchmarking (Expanse, YC P26) tested a pre-release Gemini 3.5 Pro against its resource-prediction task and reported its own specialized model outperformed 3.5 Pro and the other frontier coding models by roughly 8x on that narrow HPC workload; this is a single specialized benchmark, not a general coding ranking. Source: News – Item
Latest News
Gemini 3.5 Pro Launch Slips to July (June 24, 2026)
Business Insider reported that Google pushed the Gemini 3.5 Pro launch from June to July. Google had teased the model at I/O on May 19 with Pichai saying it would arrive "next month," but is now spending extra time gathering feedback from early testers (who have accessed it via Antigravity and LMArena) and tweaking it, including incorporating feedback that Flash 3.5 consumed tokens too quickly. The model is expected to improve on long-horizon tasks and agent powering. No price was disclosed. Sources: Businessinsider – Google 3 5 Pro July Release Tokens Ai Agents Model 2026 6 , News – Item
Free Gemini CLI Shutdown Lands on Schedule (June 18, 2026)
The June 18 cutover happened as announced: Gemini CLI and the Gemini Code Assist IDE extensions stopped serving requests for free individual users and for Google AI Pro and Ultra subscribers, who must now use Antigravity CLI. Enterprise Code Assist customers on Standard or Enterprise licenses are unaffected and keep Gemini CLI access via paid Gemini and Gemini Enterprise Agent Platform API keys. Source: Googleblog – An Important Update Transitioning Gemini Cli To Antigravity Cli
Consumer Gemini Code Assist on GitHub Sunsets (June 18 / July 17, 2026)
The consumer version of Gemini Code Assist on GitHub was deprecated June 18, 2026 (no new installations) and will fully shut down July 17, 2026, ending all code review activities. The enterprise version is not affected. The deprecation doc was last updated 2026-05-27. Source: Google – Consumer Code Review
Interactions API Reaches General Availability (June 2026)
Google promoted the Interactions API to GA as its "primary API for interacting with Gemini models and agents," with a stable schema plus Managed Agents (Antigravity agent default in a remote Linux sandbox), background execution, mixed built-in tool calling, Deep Research with two new agent versions (speed vs depth), media generation (Nano Banana 2, Lyria 3, multi-speaker TTS), Flex (50% reduction) and Priority tiers, and 55-day interaction retention on the paid tier. All documentation now defaults to Interactions API; the legacy generateContent API stays supported but frontier long-running and agentic capabilities will increasingly land only on Interactions API. Source: Google – Interactions Api General Availability
Gemini Spark Reaches macOS and Adds MCP (June 30, 2026)
Gemini Spark launched on the macOS app in beta for Google AI Ultra subscribers (18+, US), with new connected apps (Canva, Dropbox, Instacart, OpenTable, Zillow Rentals, Google Tasks, Google Keep), custom MCP support, real-time topic tracking, and a "coming soon" remote task execution feature for assigning multi-step Mac tasks from a phone. Source: Google – Gemini Spark Updates June 2026
ADK Go 2.0 and Agent Quality Flywheel (June 30, 2026)
Google shipped ADK Go 2.0, a graph-based workflow engine for multi-agent applications with built-in human-in-the-loop and dynamic orchestration, and published guidance on driving an "agent quality flywheel" from a coding agent. Sources: Googleblog – Announcing Adk Go 20 , Googleblog – Driving The Agent Quality Flywheel From Your Coding Agent
Community Signals
The 3.5 Pro slip drew pessimism about Google's coding position
The Hacker News submission for the Gemini 3.5 Pro delay (Business Insider story, 8 points) drew concern that Google is falling behind Anthropic and OpenAI specifically in coding, which the BI report called the first major enterprise use case for modern AI.
- linzhangrun: "Gemini really needs 3.5 Pro to prove itself. Feeling pessimistic about Gemini's future." News – Item
Account bans signal a harder ecosystem lock-in after the CLI cutover
A February 2026 Reddit thread surfaced on HN documented Google banning Gemini/Antigravity Pro subscription accounts used outside the approved Antigravity/Gemini-CLI tools, with affected users reporting bans even when only using opencode. This takes on added weight after the June 18 free-tier cutover formalized the move to Antigravity CLI.
- tyfon: "My pro account also got banned yesterday, but I didn't use any clawd / molt stuff like the OP. I've only been using opencode. Seems like only openai are willing to let people use their subscriptions with 3rd party tools." News – Item
Independent benchmarking against a pre-release 3.5 Pro
The Expanse (YC P26) launch post on HN (103 points) reported benchmarking a pre-release Gemini 3.5 Pro alongside Claude Opus 4.8, GPT-5.5, and Codex 5.3 on a narrow HPC resource-prediction task. This is the only public reference to 3.5 Pro being exercised by outside developers before the July launch; it is a specialized benchmark, not a general coding ranking.
- Expanse (ismaeel_bashir), in the submission text: "We benchmarked Expanse against the state of the art at the time (Gemini 3.5 pro, Claude Opus 4.8, GPT 5.5, Codex 5.3) and outperformed them by 8x." News – Item
Quiet reaction to the consumer GitHub Code Assist sunset
The HN submission "Google Is Killing 'Gemini Code Assist on GitHub'" (6 points, 0 comments at report time) received little engagement, suggesting the community had already priced in the June 18 deprecation that the May report had flagged. Source: News – Item
Enterprise Readiness
| Feature | Available? | Details |
|---|---|---|
| SSO (SAML/OIDC) | Yes | Available via Google Cloud Identity and IAM integration. Source: Google – Pricing |
| SCIM | Yes | Google Cloud supports SCIM provisioning for user management. Source: Google – Identity |
| Audit logs | Yes | Available through Google Cloud Audit Logs and usage metrics (daily active users, chat exposures, acceptance rates, lines of code accepted). Source: Google – Code Assist |
| IP indemnity | Yes | Google's generative AI indemnification covers Gemini Code Assist licensed users for copyright infringement claims, on both Standard and Enterprise. Source: Google – Protecting Customers With Generative Ai Indemnification |
| Data residency | Yes | Available via Google Cloud regional controls and VPC Service Controls. Source: Google – Code Assist |
| HIPAA | Yes | Gemini Code Assist is covered under the Google Cloud BAA for covered workloads. |
| Air-gapped / on-prem | Yes | Available via Google Distributed Cloud Air-gapped for fully disconnected deployment. Source: Google – Distributed Cloud Air Gapped |
| SLA | Yes | Governed by Google Cloud SLA terms; specific uptime guarantees vary by service tier. Source: Google – Sla |
| Admin controls (RBAC) | Yes | Granular IAM permissions, VPC Service Controls, enterprise access controls, and an admin console for license management. Source: Google – Code Assist |
Terms explained:
- VPC Service Controls - lets enterprises define a security perimeter around Google Cloud resources to prevent data exfiltration. Source: Google – Code Assist
- SOC certifications - Gemini Code Assist has achieved SOC 1/2/3, ISO/IEC 27001, 27017, 27018, and 27701 certifications. Source: Google – Code Assist
Transparency Gaps
| Gap | Details | Severity |
|---|---|---|
| Gemini 3.5 Pro price and date | Slipped to July with no confirmed launch date and no published API price. Buyers planning around it cannot budget. | High |
| Usage limits undisclosed | Gemini Code Assist for individuals (free) discloses no daily or monthly request limits. Standard and Enterprise reference "increased daily usage limits" for Enterprise but publish no numbers. Source: Google – Code Assist | High |
| Agent mode limits undisclosed | Agent mode is on both Standard and Enterprise, but Enterprise gets "increased agent usage" and "increased daily usage limits" with no specific numbers. Source: Google – Pricing | High |
| Ecosystem enforcement scope | Google bans subscription accounts used outside Antigravity/Gemini-CLI, but there is no published list of which third-party clients trigger bans or a clear policy on API-key programmatic use. Source: News – Item | High |
| Antigravity CLI feature parity | Google explicitly says there is no 1:1 feature parity with Gemini CLI at cutover; only Skills, Hooks, Subagents, and Extensions (as plugins) carry over, with no published parity matrix. Source: Googleblog – An Important Update Transitioning Gemini Cli To Antigravity Cli | Medium |
| Model parameter counts | Google does not disclose parameter counts for any Gemini 3.x model. | Medium |
| Context window usage per request | Models support a 1M token context window, but Code Assist does not disclose how much context is available per request or how context usage affects subscription billing. | Medium |
| Benchmark detail | Comparison scores versus Claude Opus 4.8, Sonnet 5, GPT-5.5, and GLM-5.2 were shown as animated GIFs at I/O, not as structured data, so exact competitor numbers are not extractable. Source: Google – Gemini 3 5 | Low |
| Per-user metering | Gemini Code Assist is per-user per-month with no token metering, so cost-per-token cannot be compared with API access or metered competitors. | Low |