GOOGLE DEEPMIND · MODEL RELEASE TRACKER
Gemini 4 Neon
Gemini 4 Neon is a released Google DeepMind model (announced 2026-11-18) positioned for fast end-to-end task completion across software engineering, knowledge work, and science; it offers a 1,000,000-token context window, multi‑modal input (text, images, audio, video), and is priced per‑token for API and app use.CURRENT SNAPSHOT5/5 DIMENSIONS WITH DATA
The dimensions that change the decision.
List pricing$3 per million input tokens ($0.15 per million cached input tokens, described as a 95% discount) and $15 per million output tokens.
Context window1,000,000‑token context window.
Supported input modalitiesDescribed as able to synthesize across text, spreadsheets, PDFs, charts, audio, images, video and to turn those inputs into structured deliverables; supports features such as Computer Use, code execution, file search, and grounding via Google Search and Maps (note: image/audio generation is not supported).
Modalities supportedSupports text, images, audio, and video; can synthesize across spreadsheets, PDFs, charts, audio and video within a single 1,000,000‑token context.
Reported benchmark rankingsReports include: #3 overall with score 61 on the Artificial Analysis Intelligence Index (xhigh thinking); #2 on the Vals Index (70.9% cited); #2 on steerability in the Agent Arena Index; #1 on Legal Research Bench and Tax Agent Bench; top results on several industry benchmarks (e.g., Surge AI, Chartography 73.5%, Agents' Last Exam ~75.8%, Zapier AutomationBench and Andon Labs Vending-Bench results cited).
Artificial Analysis Intelligence Index (xhigh)Scores 61 at the xhigh reasoning level and is reported as #3 overall on the Artificial Analysis Intelligence Index.
Availability / rolloutRolling out in the Gemini API via Google AI Studio, Gemini for Business (API and model dropdown), for paid Google AI Pro and Ultra subscribers in Google Antigravity, and for Pro and Ultra subscribers in the Gemini app.
RELEASE TIMELINE
Published, source-backed release events only.
VERIFIABLE FACTS
Every value stays attached to a source and date.
RELEASE · Release dateDEVELOPER CLAIM
Introduced/announced 2026-11-18.
CAPABILITIES · Core capability claimDEVELOPER CLAIM
Positioned as "frontier-level intelligence" for fast end-to-end task completion across complex software engineering, knowledge work, and science workflows; can turn multimodal inputs into structured deliverables.
MODALITIES · Supported input modalitiesDEVELOPER CLAIM
Described as able to synthesize across text, spreadsheets, PDFs, charts, audio, images, video and to turn those inputs into structured deliverables; supports features such as Computer Use, code execution, file search, and grounding via Google Search and Maps (note: image/audio generation is not supported).
CONTEXT WINDOW · Context windowDEVELOPER CLAIM
1,000,000‑token context window.
PRICING · List pricingDEVELOPER CLAIM
$3 per million input tokens ($0.15 per million cached input tokens, described as a 95% discount) and $15 per million output tokens.
AVAILABILITY · Availability / rolloutDEVELOPER CLAIM
Rolling out in the Gemini API via Google AI Studio, Gemini for Business (API and model dropdown), for paid Google AI Pro and Ultra subscribers in Google Antigravity, and for Pro and Ultra subscribers in the Gemini app.
BENCHMARKS · Reported benchmark rankingsDEVELOPER CLAIM
Reports include: #3 overall with score 61 on the Artificial Analysis Intelligence Index (xhigh thinking); #2 on the Vals Index (70.9% cited); #2 on steerability in the Agent Arena Index; #1 on Legal Research Bench and Tax Agent Bench; top results on several industry benchmarks (e.g., Surge AI, Chartography 73.5%, Agents' Last Exam ~75.8%, Zapier AutomationBench and Andon Labs Vending-Bench results cited).
LIMITATIONS · Generation limitationsDEVELOPER CLAIM
According to reporting, image and audio generation are not supported.
MODALITIES · Modalities supportedDEVELOPER CLAIM
Supports text, images, audio, and video; can synthesize across spreadsheets, PDFs, charts, audio and video within a single 1,000,000‑token context.
CAPABILITIES · Positioning / capability focusDEVELOPER CLAIM
Positioned as delivering "frontier‑level intelligence" with fast end‑to‑end task completion across complex software engineering, knowledge work, and science workflows.
CAPABILITIES · Throughput / speedDEVELOPER CLAIM
Reported throughput about 110 tokens per second, roughly double the reported rate for Gemini 4 Carbon in one media report.
BENCHMARKS · Artificial Analysis Intelligence Index (xhigh)DEVELOPER CLAIM
Scores 61 at the xhigh reasoning level and is reported as #3 overall on the Artificial Analysis Intelligence Index.
WHAT CHANGED
Stored passport versions, without reconstructed history.
RELEASE · Announcement− Google/DeepMind announced Gemini 4 Neon on 2026-11-18.
BENCHMARKS · Artificial Analysis Intelligence Index (xhigh)+ Scores 61 at the xhigh reasoning level and is reported as #3 overall on the Artificial Analysis Intelligence Index.
AVAILABILITY · Availability / rollout+ Rolling out in the Gemini API via Google AI Studio, Gemini for Business (API and model dropdown), for paid Google AI Pro and Ultra subscribers in Google Antigravity, and for Pro and Ultra subscribers in the Gemini app.
AVAILABILITY · Availability− Rolling out in the Gemini API in Google AI Studio, available to Gemini for business (API and model dropdown), for paid Google AI Pro and Ultra subscribers in Google Antigravity, and for Pro and Ultra subscribers in the Gemini app.
CONTEXT WINDOW · Context window and output limit− A 1,000,000-token context window; reports that it can produce up to one million output tokens.
CONTEXT WINDOW · Context window+ 1,000,000‑token context window.
CAPABILITIES · Positioning / capability focus+ Positioned as delivering "frontier‑level intelligence" with fast end‑to‑end task completion across complex software engineering, knowledge work, and science workflows.
PRICING · Pricing− Priced at $3 per million input tokens ($0.15 per million cached input tokens, a 95% discount) and $15 per million output tokens.
PRICING · List pricing+ $3 per million input tokens ($0.15 per million cached input tokens, described as a 95% discount) and $15 per million output tokens.
MODALITIES · Modalities supported+ Supports text, images, audio, and video; can synthesize across spreadsheets, PDFs, charts, audio and video within a single 1,000,000‑token context.
RELEASE · Release date+ Introduced/announced 2026-11-18.
CAPABILITIES · Throughput / speed+ Reported throughput about 110 tokens per second, roughly double the reported rate for Gemini 4 Carbon in one media report.