RELEASE · MODELS · #3878
Google DeepMind launches Gemini 4 Neon — a 1M‑token multimodal frontier model
Google DeepMind announced Gemini 4 Neon, a "frontier-level" multimodal model with a 1,000,000‑token context window. Neon is available via the Gemini API, Google AI Studio, Google Antigravity, Gemini for business, and paid Gemini app subscriptions; Google published pricing ($3 per million input tokens, $0.15 per million cached input tokens, $15 per million output tokens) and highlighted high rankings on multiple internal and external benchmarks (e.g., #3 on the Artificial Analysis Intelligence Index at xhigh with score 61, #2 on the Vals Index, and top placements on several coding, finance, legal and agent benchmarks).
KEY POINTS
- Google DeepMind announced Gemini 4 Neon, a "frontier-level" multimodal model with a 1,000,000‑token context window.
- Neon is available via the Gemini API, Google AI Studio, Google Antigravity, Gemini for business, and paid Gemini app subscriptions; Google published pricing ($3 per million input tokens, $0.15 per million cached input tokens, $15 per million output tokens) and highlighted high rankings on multiple internal and external benchmarks (e.g., #3 on the Artificial Analysis Intelligence Index at xhigh with score 61, #2 on the Vals Index, and top placements on several coding, finance, legal and agent benchmarks).
- The release signals Google’s push to ship a fast, long‑context, agent‑optimized foundation model for enterprise and developer workflows and may shift competitive dynamics and product integrations across AI tooling.
DECISION BRIEF
CONFIDENCE
WHAT CHANGED
Google DeepMind announced Gemini 4 Neon, a “frontier‑level” multimodal model with a 1,000,000‑token context window. Google published per‑token pricing ($3 per 1M input tokens, $0.15 per 1M cached input tokens, $15 per 1M output tokens) and said Neon is available via the Gemini API, Google AI Studio, Gemini for business (API and model dropdown), Google Antigravity for paid Pro/Ultra subscribers, and the Gemini app for Pro/Ultra subscribers. Google reported multiple benchmark placements and internal test results.
WHY NOW
The announcement packages Neon as a fast, long‑context, agent‑oriented model for complex software engineering, knowledge work, and science workflows and is being rolled out immediately across Google’s developer and paid product surfaces. That combination of claimed capability plus immediate availability and published pricing makes the model directly usable by developers and enterprise customers today.
WHO IS AFFECTED
Directly affected parties named in the announcement: developers and integrators using the Gemini API and Google AI Studio; businesses using Gemini for business; paid Google Antigravity and Gemini app Pro/Ultra subscribers; and enterprise users working with multi‑format inputs (spreadsheets, PDFs, charts, audio, video) per Google’s description.
CONFIRMED
Explicit, source‑backed facts in the announcement excerpts: - The product name is Gemini 4 Neon. - The model has a 1,000,000‑token context window. - Published pricing: $3 per million input tokens, $0.15 per million cached input tokens, $15 per million output tokens. - Availability: rolling out in the Gemini API, Google AI Studio, Gemini for business (via API and model dropdown), Google Antigravity for paid Pro and Ultra subscribers, and the Gemini app for Pro and Ultra subscribers. - Google describes Neon as targeted at complex software engineering, knowledge work, and science workflows and able to synthesize spreadsheets, PDFs, charts, audio, and video in a single 1M‑token context. - Benchmark and test results reported by Google/DeepMind in the excerpts: #3 on the Artificial Analysis Intelligence Index at xhigh (score 61), #2 on the Vals Index (70.9% cited in the excerpt), #2 on steerability in the Agent Arena Index, #1 on Legal Research Bench and Tax Agent Bench, top placements on multiple coding, finance, legal and agent benchmarks, #1 on Zapier’s AutomationBench (56.3%), #1 on many Surge AI benchmarks across finance/accounting/spreadsheet modeling, #1 on Andon Labs' Vending‑Bench 2 (result reported as $16,843), #2 on Arena's Agent Arena Index globally (score 1336 across 4,200+ agent sessions), 73.5% on Chartography, 75.8% on Agents' Last Exam v1.1, and Google reports its lowest fabrication rate yet in internal Google Antigravity testing.
UNCERTAIN
Missing or conflicting evidence in the provided excerpts: - The excerpts do not disclose model architecture details (parameter count, training data, training compute). - Latency, throughput, per‑request or per‑customer rate limits, regional availability, and SLA details are not provided. - Long‑term pricing commitments, enterprise licensing terms, discounts for high volume or education/nonprofit are not described. - The excerpts mix Google/DeepMind reporting and named third‑party benchmarks; some tests are explicitly internal (Google Antigravity) while others are third‑party (e.g., Surge AI, Zapier, Andon Labs, Arena), but the excerpts do not provide full methodology or independent verification details for each benchmark. - Real‑world adoption, integration complexity, and cost‑effectiveness in customer deployments are not reported in the excerpts. - Any safety, content‑policy, or moderation behavior beyond the noted reduction in fabrication rate is not documented in these excerpts.
WHAT TO WATCH
Concrete observable signals to monitor (using sources named in the excerpts where relevant): - Publication of a technical report or model card with architecture, training data, and evaluation methodology. - Third‑party benchmark releases and independent evaluations on indexes named (Artificial Analysis Intelligence Index, Vals Index, Agent Arena Index, Surge AI, Zapier AutomationBench, Andon Labs, Arena, Chartography, Agents' Last Exam). - Detailed Google product documentation or API release notes showing rate limits, latency, regional availability, and enterprise terms. - Broader availability changes (movement out of paid Pro/Ultra or expansion to other Google product tiers). - Customer case studies, adoption announcements, or integration guides from Google or enterprise partners. - Any updates from Google about safety, fabrication rates, or mitigation measures with data beyond the internal Antigravity statement.
WHY IT MATTERS
The release signals Google’s push to ship a fast, long‑context, agent‑optimized foundation model for enterprise and developer workflows and may shift competitive dynamics and product integrations across AI tooling.
EVIDENCE MAP
4Editorial claims linked to specific sources, with support, contradiction and context shown separately.
Gemini 4 Neon has a 1,000,000‑token context window.
SUPPORTEDChecked 2026-11-18 · 2 supporting
Google published Neon pricing: $3 per million input tokens, $0.15 per million cached input tokens (95% discount), and $15 per million output tokens.
SUPPORTEDChecked 2026-11-18 · 2 supporting
Gemini 4 Neon is rolling out via the Gemini API, Google AI Studio, Gemini for Business, Google Antigravity (paid Pro/Ultra), and the Gemini app (Pro/Ultra subscribers).
SUPPORTEDChecked 2026-11-18 · 2 supporting
Article says Neon is rolling out in the Gemini API, Google AI Studio, Gemini for business, Google Antigravity for paid Pro/Ultra subscribers, and the Gemini app for Pro/Ultra subscribers.
SUPPORTS Google DeepMindMEDIA · 2026-11-18DeepMind blog provides the same rollout/availability details.
Neon ranks #3 overall with a score of 61 on the Artificial Analysis Intelligence Index at the xhigh thinking level.
SUPPORTEDChecked 2026-11-18 · 2 supporting
SOURCES & TIMELINE
2Bringing frontier-level intelligence with fast end-to-end task completion across complex software engineering, knowledge work, and science workflows. SVP, Google DeepMind and Chief AI Architect, Google Today, we’re introducing Gemini 4 Neon — bringing frontier-level intelligence with fast end-to-end task completion across complex software engineering, knowledge work, and science workflows. Neon ranks ahead of comp…
Bringing frontier-level intelligence with fast end-to-end task completion across complex software engineering, knowledge work, and science workflows. SVP, Google DeepMind and Chief AI Architect, Google Your browser does not support the audio element. Today, we’re introducing Gemini 4 Neon — bringing frontier-level intelligence with fast end-to-end task completion across complex software engineering, knowledge work…