Jornal Em Destaque reports that Google released Gemini 3.8 Flash as its latest Flash AI model with a standard "workhorse" variant for software development and agentic work, and a cybersecurity-focused sibling called Gemini 3.8 Flash Cyber tuned to detect and help fix software vulnerabilities. The model delivers Google' Developer API access is available at an introductory price of $0.75 per million input tokens and $3.75 per million output tokens through the end of 2026, after which regular pricing rises to $1.50 and $7.50 respectively. The model also improves computer-use performance on the OSWorld-2.0 test over Gemini 3.7 Flash, tho

The eesel AI review of Gemini 3.8 Flash confirms the model ships as gemini-3.8-flash with unchanged generation-over-generation specifications: text, image, audio, video, and PDF inputs with text outputs, a 1M token context window, a 64K output ceiling, function calling, search-as-a-tool, and computer-use capability. De The review highlights Gemini 3.8 Flash Cyber as the more impressive launch, a security variant with deliberately looser mitigations for cyber work gated to trusted defenders through Google's new Fairwind Program. Pricing is identical to 3.7 Flash at $0.75/$3.75 per million tokens through December 31, 2026, then doubles

The Register reports that Google announced Gemini 3.8 Flash on September 2, 2026 as its fourth Flash model in as many months, crediting the release to named Google DeepMind sources Tulsee Doshi (senior director of product management) and Raluca Ada Popa (Gemini Security Lead). Their blog post states that "Gemini 3.8 Fl Introductory API pricing is $0.75 per million input tokens and $3.75 per million output tokens, doubling in early 2027 once the promotion expires. The Register notes that the rapid iteration has helped Google's Flash models stay competitive on intelligence benchmarks even as open-weight models from Chinese AI labs have

Ap7i's coverage of the September 2, 2026 Gemini 3.8 Flash launch frames the model as engineered for long-horizon software engineering, autonomous tool orchestration, and multi-step domain reasoning while preserving 3.7-tier introductory pricing of $0.75 per million input tokens and $3.75 per million output tokens. The For autonomous agent harnesses including Google Antigravity and GitHub Copilot, the diligence tuning is positioned to translate into fewer abandoned runs and more reliable multi-step completions. The piece also flags that the model may consume more tokens on complex problems due to deeper deliberation, trading higher i

Google released Gemini 3.8 Flash on September 2, 2026, its third Flash-tier model in roughly six weeks, with a cybersecurity-tuned sibling called Gemini 3.8 Flash Cyber built on the same foundation but specialized for vulnerability detection and mitigation. Ars Technica reports that the standard variant is pitched as a For developers, API access is available at an introductory rate of $0.75 per million input tokens and $3.75 per million output tokens through the end of 2026, after which regular pricing doubles to $1.50 and $7.50 respectively. Ars Technica notes Google has not shipped a Gemini Pro frontier model since early 2026 and f

Alphacorp's launch guide confirms Google DeepMind released Gemini 3.8 Flash on September 2, 2026 at $0.75 per million input tokens and $3.75 per million output tokens, unchanged from 3.7 Flash, with a roughly six-point gain on long-horizon coding and tuning for multi-step agent tasks. Published benchmark figures includ Standard-tier pricing of $0.75 in and $3.75 out per million tokens holds through December 31, 2026, after which it doubles for both input and output tokens. The guide frames the launch around migration considerations for production traffic, including workload fit for long-horizon coding and agentic tasks, and flags tha

The llm-stats.com model page for Gemini 3.8 Flash aggregates benchmark scores sourced from deepmind.google, including GDPval-AA at 1545/3000 (Artificial Analysis public leaderboard), Terminal-Bench 2.1 at 0.89/1 (Terminus 2 default agent harness), LVBench at 0.87/1 across 1024 frames with no tools, and CharXiv-R at 0.8 The page presents Gemini 3.8 Flash's performance across datasets with each benchmark's methodology described in-line, including Elo scoring for GDPval-AA v2, autonomous end-to-end task evaluation for Terminal-Bench 2.1, extreme long-video understanding up to two hours for LVBench, and scientific-chart reasoning for Cha

Fello AI's technical analysis confirms Gemini 3.8 Flash reached general availability on September 2, 2026, three weeks after Gemini 3.7 Flash and six weeks after 3.6 Flash, at $0.75 per million input tokens and $3.75 per million output tokens. It scores 73.7% on DeepSWE v1.1 per Google's evaluation PDF and wins 8 of 14 The review notes that the largest single gain over 3.7 Flash is BioMysteryBench Human Difficult (43.5% to 56.5%), followed by Terminal-bench 4.0 (11.2% to 19.1%). API pricing is a promotion that expires December 31, 2026, after which rates double to $1.50 and $7.50, and context caching doubles alongside. The model is t