model release

Google releases Gemini 3.5 Flash with autonomous coding and agent capabilities, claims 4x speed boost

TL;DR

Google released Gemini 3.5 Flash, positioning it as an agent-first model designed for autonomous coding and multi-hour workflows. The company claims the model outperforms its 3.1 Pro predecessor on coding and agentic benchmarks while running 4x faster than competing frontier models, with an optimized version achieving 12x speed gains.

3 min read
0

Gemini 3.5 Flash at a glance

Context window1049K tokens
Input$1.5/1M tokens
Output$9/1M tokens

How Gemini 3.5 Flash ranks now

BenchmarkScoreRankSource
Arena Elo1478 Elo26 of 141LMArena, Nov 2026 ↗
GPQA92.2%25 of 164Artificial Analysis, Jul 2026 ↗
Speed (tok/s)167 tok/s17 of 103Artificial Analysis, Nov 2026 ↗

Rank is against every model in our database with a sourced score on that benchmark, each counted once with its best score.

All Gemini 3.5 Flash specs and scores →

Google releases Gemini 3.5 Flash with autonomous coding and agent capabilities, claims 4x speed boost

Google released Gemini 3.5 Flash at its I/O developer conference on Tuesday, positioning the model as purpose-built for autonomous agents rather than conversational AI. According to the company, the model can independently execute coding pipelines, manage research projects, and in internal demonstrations, built an operating system from scratch.

Performance claims

"It outperforms our latest frontier model, 3.1 Pro, on nearly all the benchmarks," said Koray Kavukcuoglu, DeepMind's chief technologist, including coding, agentic tasks, and multimodal reasoning. Google claims the model runs 4x faster than other frontier models, with an optimized version achieving 12x speed improvements "with the same quality."

Specific benchmark scores were not disclosed. Pricing details were not announced.

Agent-first architecture

The speed improvements target multi-agent workflows where multiple AI instances run simultaneously on long-duration tasks. Google demonstrated the model at I/O spawning separate agents to build components of a full operating system inside Antigravity, the company's agent development platform.

Kavukcuoglu said Flash 3.5 was co-developed with Antigravity to create a "native environment where they can live, work, and execute." Google released Antigravity 2.0 alongside the model, describing it as a standalone desktop IDE designed for agent-first development.

The model can run autonomously for multiple hours. Tulsee Doshi, Google's senior director and head of product, said it pauses to request user input at decision points requiring human judgment.

Multi-model orchestration

Google plans to release Gemini 3.5 Pro, which will work in tandem with Flash. "3.5 Pro becomes your orchestrator, your planner, and then it actually can leverage Flash to be the various sub-agents," Doshi said. The Pro model will handle reasoning-intensive tasks while Flash executes tool use and sub-tasks.

Deployment and access

Gemini 3.5 Flash is now the default model in the Gemini app and AI Mode in Search globally. It's available through Antigravity, the Gemini API, and Gemini Enterprise. The model will also power Gemini Spark, Google's 24/7 personal AI agent for consumer digital life management.

Google announced agentic capabilities coming to Search, allowing users to create and manage AI agents directly on the platform.

Safety considerations

The deployment comes as Google faces a lawsuit after a user died by suicide following weeks of interactions with Gemini in 2025. Google says Gemini 3.5 includes strengthened cyber and CBRN (Chemical, Biological, Radiological, and Nuclear) safeguards and is "better calibrated to engage with sensitive questions rather than refuse them outright."

What this means

Google is repositioning its AI strategy around autonomous agents capable of multi-hour workflows rather than single-turn conversations. The speed claims—if validated—would represent a significant advantage for agentic applications where latency compounds across multiple model calls. However, without disclosed benchmarks, context window size, or pricing, developers lack the data needed to evaluate the model against Anthropic's Claude 3.5 Sonnet or OpenAI's o1 for similar use cases. The co-development of Flash with Antigravity suggests Google is betting on vertical integration between model and tooling as a competitive moat.

Related Articles

model release

RadixArk releases DaisyRose-27B, lifting Terminal-Bench 4.1 pass@1 from 5.6% to 12.0%

RadixArk released DaisyRose-27B, its first model, built on Qwen3.8-27B and trained with reinforcement learning on 1,200 curated agent tasks. According to RadixArk, it raises Terminal Bench 4.1 pass@1 from 5.6% to 12.0% and scores higher than the base model on all seven reported benchmarks. Weights ship under Apache 2.0 in BF16, FP8 and NVFP4.

model release

Google releases Genie 4 to AI Ultra subscribers with real-time prompt editing and native audio

Google is rolling out Genie 4, which it calls its most realistic and interactive world model yet, to eligible Google AI Ultra subscribers in Project Genie. The update adds real-time native audio, mid-session text prompting, a world-building agent and a community gallery. Genie 3 also gains sessions of up to 3 minutes.

model release

OpenAI's GPT-6.2 Astra ranks second on Artificial Analysis index at 62, behind Claude Fable 5.2 (65)

OpenAI released GPT-6.2 Astra and GPT-6.2 Sol. Artificial Analysis ranks Astra second in its Intelligence Index with 62 points, behind Claude Fable 5.2 at 65, and finds coding performance flat versus the previous generation. Astra's API price is $8 per 1M input tokens and $40 per 1M output tokens.

model release

Microsoft releases Aion 1.0 Instruct, a 2.3B on-device SLM with 32K context under MIT license

Microsoft's Windows Applied Sciences Group has published Aion 1.0 Instruct, a roughly 2.29B-parameter instruction-tuned small language model with a 32,768-token context window under the MIT license. A further optimized version is slated to become the Windows OS SLM by the end of January 2027.

Comments

Loading...