Article · 2026-08-13 · 4 moments

llm-gemini 0.33

Release: llm-gemini 0.33 It's been a while since the last llm-gemini release. This version of the plugin adds support for today's Gemini 3.7 Flash release, plus gemini-3.6-flash, gemini-3.5-flash-lite and two embedding models gemini-embedding-2 and gemini-embedding-001. The plugin is also upgraded for compatibility with LLM 0.32, which means you can now see reasoning traces and you can also enable server-side tools using this pattern: llm -m gemini-3.7-flash -T CodeExecution \ 'use ✦ AI generated

Read original ↗

HOST Simon Willison

Example

Gemini 3.7 Flash can generate high-quality images (here, pelicans riding bicycles) and the 'minimal' thinking-effort option from 3.6 Flash has been removed in 3.7.

The author had Gemini 3.7 Flash generate pelicans riding bicycles at high, medium, and low thinking efforts, noting that the 'minimal' option from 3.6 Flash has been removed in 3.7 and praising the high-effort result.

transcript

Simon Willison: I had Gemini 3.7 Flash draw me some pelicans riding bicycles at high, medium, and low thinking efforts (minimal, which was an option in 3.6 Flash, has been removed in 3.7.) Here's the high level one, which is pretty great:

gives example · 1

Fact◆

llm-gemini 0.33 adds support for Gemini 3.7 Flash, gemini-3.6-flash, gemini-3.5-flash-lite, and two embedding models (gemini-embedding-2 and gemini-embedding-001) in a single release.

The release notes announce llm-gemini 0.33, which adds support for today's Gemini 3.7 Flash release, plus several other Gemini chat and embedding models.

transcript

Simon Willison: It's been a while since the last llm-gemini release. This version of the plugin adds support for today's Gemini 3.7 Flash release, plus gemini-3.6-flash, gemini-3.5-flash-lite and two embedding models gemini-embedding-2 and gemini-embedding-001.

Mechanism◆

llm-gemini 0.33 is upgraded for compatibility with LLM 0.32, enabling users to see reasoning traces and enable server-side tools via the -T CodeExecution pattern.

The plugin is upgraded to work with LLM 0.32, which lets users see reasoning traces and enable server-side tools using the -T CodeExecution flag, demonstrated with a factorial calculation command.

transcript

Simon Willison: The plugin is also upgraded for compatibility with LLM 0.32, which means you can now see reasoning traces and you can also enable server-side tools using this pattern: llm -m gemini-3.7-flash -T CodeExecution \ 'use python to calculate (factorial of 13) * 3'

Context

The generated pelican SVG renders differently across browsers because Safari is more tolerant of empty SVG <filter> elements than Firefox and Chrome, leading to the pelican disappearing in the latter two.

The author notes that the pelican image, created as an SVG, renders fully in Safari but loses the pelican in Firefox and Chrome because Safari tolerates empty SVG <filter> elements more permissively than the other browsers.

transcript

Simon Willison: One catch though: the pelican I showed here was rendered with Safari. Both Firefox and Chrome render it differently, due to Safari being more tolerant of empty SVG <filter> elements than those other two browsers. They still display the bicycle, but the pelican is missing entirely!

Highlight slides

llm-gemini 0.33 adds Gemini 3.7 Flash and more✦ from: llm-gemini 0.33 adds support for Gemini 3.7 Flash, gemini-3.6-flash, gemini-3.5-flash-lite, and two embedding models (gemini-embedding-2 and gemini-embedding-001) in a single release.llm-gemini 0.33 upgrade for LLM 0.32 compatibility✦ from: llm-gemini 0.33 is upgraded for compatibility with LLM 0.32, enabling users to see reasoning traces and enable server-side tools via the -T CodeExecution pattern.

Related episodes

Video · The TWIML AI Podcast

Gemini 3.7 Flash can generate high-quality images (here, pelicans riding bicycles) and the 'minimal' thinking-effort option from 3.6 Flash has been removed in 3.7.

llm-gemini 0.33 adds support for Gemini 3.7 Flash, gemini-3.6-flash, gemini-3.5-flash-lite, and two embedding models (gemini-embedding-2 and gemini-embedding-001) in a single release.

llm-gemini 0.33 is upgraded for compatibility with LLM 0.32, enabling users to see reasoning traces and enable server-side tools via the -T CodeExecution pattern.

The generated pelican SVG renders differently across browsers because Safari is more tolerant of empty SVG <filter> elements than Firefox and Chrome, leading to the pelican disappearing in the latter two.

The Race to Production-Grade Diffusion LLMs [Stefano Ermon] - 764

[AINews] Lilian Weng summarizes 35 papers on Harness Engineering for RSI

Why Image Generation Needs More Than Bigger Models