Google and Meta Released Rival AI Models Hours Apart: Who Leads?

Google and Meta launched their newest frontier AI fashions inside hours of one another on Wednesday. Google shipped Gemini 3.8 Flash alongside a cybersecurity variant, whereas Meta pushed out Muse Spark 1.3.

The two launches invite a direct comparability. Independent testing by Artificial Analysis splits the end result. Meta leads on agentic information work and scientific reasoning, whereas Google holds an edge in factual recall and terminal coding.

Two Frontier Releases Land on the Same Day

Gemini 3.8 Flash is Google’s third Flash release in six weeks. The mannequin prices $0.75 per 1 million enter tokens and $3.75 per 1 million output tokens.

Follow us on X to get the newest information because it occurs

Today we’re introducing a brand new 3.8 Flash mannequin, our third Flash launch in simply 6 wks.

It delivers vital leaps from 3.7 Flash throughout software program engineering, agentic duties, and multi-step reasoning. On DeepSWE v1.1, it outperforms most bigger frontier fashions in autonomously… pic.twitter.com/MB5seliluh

— Sundar Pichai (@sundarpichai) September 2, 2026

That introductory charge runs by way of December 31, 2026. Prices then double to $1.50 and $7.50 per 1 million tokens.

Google paired the overall mannequin with Gemini 3.8 Flash Cyber. The mannequin is out there to a set of trusted defenders. It scored 86.2% on CyberGymnasium, a benchmark for locating vulnerabilities.

The cyber variant additionally reached 47.2% on CWE-Bench, a patching benchmark. Google mentioned the mannequin produced 2.6 instances extra right patches for Chrome vulnerabilities than bigger industrial fashions.

Access sits behind the Fairwind Program, which limits the mannequin to authorities authorities and vital infrastructure operators. OpenAI drew an identical boundary a day earlier round Astra, the primary mannequin it rated at a critical cybersecurity threshold.

Meanwhile, Meta rolled out Muse Spark 1.3 by way of Muse Code and the Meta Model API. Company engineers measured roughly 20% fewer software calls than model 1.2.

Muse Spark 1.3 is rolling out immediately with frontier efficiency nearly too low-cost to meter. This is the largest soar we have made to this point on coding and agentic work. Try it in Muse Code and our API.

Next up 🍉 and Muse Spark open weights releases coming quickly. pic.twitter.com/XQQEDEJGD7

— Mark Zuckerberg (@finkd) September 2, 2026

Gemini 3.8 Flash vs Muse Spark 1.3: Independent Benchmarks Split the Result

Artificial Analysis tested the fashions, and the results show how they rank. Muse Spark 1.3 in max mode scored 1,754 Elo on GDPval-AA v2. Gemini 3.8 Flash (high) returned 1,545.

Meta additionally led the Sierra Research banking agent take a look at (52.4% to 44.9%) and CritPt physics reasoning. Gemini 3.8 Flash led Terminal-Bench 2.1 at 87.6%, AA-LCR lengthy context at 81%, and AA-Omniscience accuracy at 55%.

Gemini 3.8 Flash posted the very best GPQA Diamond rating among the many fashions examined, at 95%. The two completed inside some extent of one another on Humanity’s Last Exam.

Meta’s prime scorer doesn’t ship immediately. The firm mentioned max reasoning will arrive as soon as additional security testing is full, leaving xhigh because the accessible variant.

That model scored 61 on the Artificial Analysis Intelligence Index, 4 factors above Muse Spark 1.2. It trails Claude Fable 5.1 at 66 and Claude Opus 5 at 63.

More releases are already queued. Elon Musk has mentioned Grok 4.7 arrives shortly, which might place 4 frontier launches inside a fortnight.

Subscribe to our YouTube channel to look at leaders and journalists present knowledgeable insights

The submit Google and Meta Released Rival AI Models Hours Apart: Who Leads? appeared first on BeInCrypto.