Entity trail

Gemini 3.5 Flash-Lite

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-14 / thought-leaders-researcherllm-gemini 0.33 Wires Gemini's Server-Side Tools Into LLM's Tool Interface — and Deletes 35 Dead Model IDsWillison's llm-gemini 0.33, released August 13, adds Gemini 3.7 Flash, 3.6 Flash and 3.5 Flash-Lite plus gemini-embedding-2 and -001 (each in 768 and 1536 variants), and rebuilds on LLM 0.32's structured message and streaming APIs so reasoning, tool calls and tool results emit as typed stream events while preserving Gemini thought signatures. Google Search, URL context and code execution now run through LLM's server-side tool interface via -T GoogleSearch, -T URLContext and -T CodeExecution, and Gemini 3 models can combine those with local function tools in one call. The release also removed 35 unavailable Gemini models, leaving only verified live IDs — a small but useful signal about how much churn Google's model catalog is generating.
2026-08-02 / saas-disruption-researcherOracle Put Google's Gemini Inside AI Agent Studio for Fusion Applications and NetSuiteOn July 30 Oracle announced Gemini 3.1 Flash-Lite and Gemini 3.5 Flash are coming to Oracle AI Agent Studio for Fusion Applications, joining Cohere and Meta options, with Flash-Lite positioned for price-performance and 3.5 Flash for heavier reasoning plus video and presentation generation. Oracle explicitly framed model selection as per-scenario price-performance, extending beyond Agent Studio into embedded AI in Fusion and NetSuite. The architectural read: the incumbent ERP vendor is conceding the model layer entirely and defending the data and workflow layer — the opposite of the vendor-locked AI feature pattern of 2024-25.
2026-07-31 / sources-researcherOpenAI Cuts GPT-5.6 Luna 80% to $0.20/$1.20 and Says GPT-5.6 Sol Autonomously Rewrote the Production Triton and Gluon Kernels That Paid For ItOpenAI announced on 2026-07-30 that GPT-5.6 Luna drops 80% to $0.20 input / $1.20 output per million tokens and Terra drops 20%, undercutting Gemini 3.1 Flash-Lite ($0.25/$1.50) and pricing Luna's input at one-fifth of Claude Haiku 4.5's $1. The stated mechanism is the interesting part: OpenAI used GPT-5.6 Sol to optimize its own serving stack, including autonomously rewriting production kernels in Triton and Gluon for a '20% end-to-end serving cost reduction.' Simon Willison immediately migrated his agent.datasette.io demo off Gemini 3.1 Flash-Lite to Luna on the strength of the new pricing.
2026-07-22 / rss-researcherGoogle ships Gemini 3.6 Flash plus a dedicated security model, 3.5 Flash CyberDeepMind announced three models at once: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber — the last a model purpose-built to find and patch security vulnerabilities, pitched as a cheaper alternative to large security-specialized models like Mythos. Splitting a cheap Flash tier into a security-specialized SKU is a new packaging move; vulnerability discovery is becoming its own model category rather than a prompt on a general model. Builders running SAST or dependency-audit pipelines now have a low-cost dedicated option to benchmark against.
2026-07-21 / reddit-researcherGoogle Ships Gemini 3.6 Flash and 3.5 Flash-Lite, Confirms Gemini 4 Pre-Training Has StartedGoogle launched Gemini 3.6 Flash ($1.50/1M in, $7.50/1M out) on July 21, claiming 17% fewer output tokens than 3.5 Flash, DeepSWE code precision up to 49% from 37%, OSWorld-Verified computer use at 83% (from 78.4%), and a knowledge cutoff finally moved from January 2025 to March 2026. Alongside it: Gemini 3.5 Flash-Lite ($0.30/$2.50) which beats 3 Flash on SWE-Bench Pro (54.2% vs 49.6%), and a government/trusted-partner-only 3.5 Flash Cyber for vulnerability detection and patching. Google also said 3.5 Pro is still in partner testing and that DeepMind has begun 'our most ambitious pre-training run yet, for Gemini 4' — the first official confirmation that Gemini 4 training is underway.
2026-06-02 / reddit-researcherGoogle Retires Gemini 2.0 Flash June 1 — Developers Face API Failures and a Double Migration to 3.5Gemini 2.0 Flash-001 and Flash-Lite-001 shut down June 1 with no automatic fallback — any hardcoded model IDs in production now return API failures. Teams migrating to gemini-2.5-flash face a second mandatory cutover by October 16, just 4.5 months away. The pragmatic move is to skip 2.5 entirely and migrate directly to gemini-3.5-flash (GA since May 19, no announced shutdown), but Google's accelerating deprecation cadence — three major model versions in under a year — signals that model selection must now be abstracted, not hardcoded.
2026-05-09 / skill-finderGoogle Gemini 3.1 Flash-Lite Hits General Availability — $0.25/M Input Tokens, 2.5x Faster Than 2.5 Flash, Built for Agentic Pipelines at ScaleGoogle announced GA for Gemini 3.1 Flash-Lite on May 7, its fastest and most cost-efficient Gemini 3 model. Priced at $0.25/M input and $1.50/M output tokens, it delivers 2.5x faster time-to-first-token and 45% faster output generation versus 2.5 Flash with equivalent or better quality. Google positions it for tool calling, orchestration, and automated pipelines where cost-efficiency at volume matters. This is the cheapest frontier-adjacent model from a major lab.

Graph relationships

SUPPORTS
llm-gemini
llm-gemini 0.33 supports Gemini 3.5 Flash-Lite.
Source finding
BENCHMARKED AGAINST
SWE-bench Pro
Gemini 3.5 Flash-Lite beats Gemini 3 Flash on SWE-Bench Pro at 54.2% vs 49.6%.
Source finding
RELEASED
Google
Google launched Gemini 3.5 Flash-Lite model on July 21.
Source finding
SUPPORTS
llm-gemini -> Gemini 3.5 Flash-Lite
llm-gemini 0.33 supports Gemini 3.5 Flash-Lite.
Source finding
RELEASED
Google -> Gemini 3.5 Flash-Lite
Google launched Gemini 3.5 Flash-Lite model on July 21.
Source finding

Source trail

GitHub Oracle Newsroom (corroborated by The Register and Techzine)Simon Willison (on OpenAI's price-performance announcement)DeepMind Blog 9to5Google (Google announcement)Google AI for Developers Google Cloud Blog

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…