❌

Vue normale

Il y a de nouveaux articles disponibles, cliquez pour rafraîchir la page.
Aujourd’hui — 18 septembre 2026Flux principal

Gemini Live audio

✇Simon Willison's Weblog
Par : Simon Willison
16 septembre 2026 à 00:47

Tool: Gemini Live audio

Google released Gemini 3.8 Live and 3.8 Live Extended Thinking today - two new speech-to-speech models that are a similar shape to OpenAI's GPT-Live family.

I pointed GPT-6 Astra Extra High at the documentation and had it build me this web UI for trying out the new models. You can select a model and voice preset, enter an optional system prompt and then start a voice conversation through your browser, including the ability to interrupt the model while it is talking.

Screenshot of a voice chat web interface with a transcript. Top buttons: Start session, End session, Mute mic, plus a Mic level meter and a timer showing 0:33. Status: Listening. Use headphones to reduce echo. Starting a session asks for microphone access. Transcript (with Download transcript and Clear buttons). Gemini: Yes, it's working perfectly. I can hear you clearly. How can I help you today? You: Okay, this is pretty good. Tell me some interesting facts about the California brown pelican. Gemini: They're famous for their spectacular plunge-dives to catch fish and have huge throat pouches that can hold up to three gallons of water and fish. They also made a great recovery after being endangered — marked as Interrupted. You: No, tell me different facts actually. Gemini: They nest in colonies on offshore islands and can hold up to three gallons of water in their throat pouches. At the bottom is a text input reading Or type a message… with a Send button, and the note: Sending a message interrupts the current response. Transcripts may include speech interrupted before playback.

The implementation uses no libraries. It connects to the wss://generativelanguage.googleapis.com/ws/google.ai.generativelanguage.v1alpha.GenerativeService.BidiGenerateContent?key=... WebSocket endpoint and uses a Web Audio API AudioContext for both capture and playback.

Here's the Gemini Live tutorial for getting started with that WebSockets API.

Tags: google, tools, websockets, generative-ai, llms, gemini, llm-release, speech-to-text

À partir d’avant-hierFlux principal

Introducing GPT-5.2-Codex

✇Simon Willison's Weblog
Par : Simon Willison
19 décembre 2025 à 06:21

Introducing GPT-5.2-Codex

The latest in OpenAI's Codex family of models (not the same thing as their Codex CLI or Codex Cloud coding agent tools).

GPT‑5.2-Codex is a version of GPT‑5.2⁠ further optimized for agentic coding in Codex, including improvements on long-horizon work through context compaction, stronger performance on large code changes like refactors and migrations, improved performance in Windows environments, and significantly stronger cybersecurity capabilities.

As with some previous Codex models this one is available via their Codex coding agents now and will be coming to the API "in the coming weeks". Unlike previous models there's a new invite-only preview process for vetted cybersecurity professionals for "more permissive models".

I've been very impressed recently with GPT 5.2's ability to tackle multi-hour agentic coding challenges. 5.2 Codex scores 64% on the Terminal-Bench 2.0 benchmark that GPT-5.2 scored 62.2% on. I'm not sure how concrete that 1.8% improvement will be!

I didn't hack API access together this time (see previous attempts), instead opting to just ask Codex CLI to "Generate an SVG of a pelican riding a bicycle" while running the new model (effort medium). Here's the transcript in my new Codex CLI timeline viewer, and here's the pelican it drew:

Alt text by GPT-5.2-Codex: A minimalist illustration of a white pelican with a large orange beak riding a teal bicycle across a sandy strip of ground. The pelican leans forward as if pedaling, its wings tucked back and legs reaching toward the pedals. Simple gray motion lines trail behind it, and a pale yellow sun sits in the top‑right against a warm beige sky.

Tags: ai, openai, generative-ai, llms, pelican-riding-a-bicycle, llm-release, codex-cli, gpt-codex

Gemini 3 Flash

✇Simon Willison's Weblog
Par : Simon Willison
17 décembre 2025 à 23:44

It continues to be a busy December, if not quite as busy as last year. Today's big news is Gemini 3 Flash, the latest in Google's "Flash" line of faster and less expensive models.

Google are emphasizing the comparison between the new Flash and their previous generation's top model Gemini 2.5 Pro:

Building on 3 Pro’s strong multimodal, coding and agentic features, 3 Flash offers powerful performance at less than a quarter the cost of 3 Pro, along with higher rate limits. The new 3 Flash model surpasses 2.5 Pro across many benchmarks while delivering faster speeds.

Gemini 3 Flash's characteristics are almost identical to Gemini 3 Pro: it accepts text, image, video, audio, and PDF, outputs only text, handles 1,048,576 maximum input tokens and up to 65,536 output tokens, and has the same knowledge cut-off date of January 2025 (also shared with the Gemini 2.5 series).

The benchmarks look good. The cost is appealing: 1/4 the price of Gemini 3 Pro ≤200k and 1/8 the price of Gemini 3 Pro >200k, and it's nice not to have a price increase for the new Flash at larger token lengths.

It's a little more expensive than previous Flash models - Gemini 2.5 Flash was $0.30/million input tokens and $2.50/million on output, Gemini 3 Flash is $0.50/million and $3/million respectively.

Google claim it may still end up cheaper though, due to more efficient output token usage:

> Gemini 3 Flash is able to modulate how much it thinks. It may think longer for more complex use cases, but it also uses 30% fewer tokens on average than 2.5 Pro.

Here's a more extensive price comparison on my llm-prices.com site.

Generating some SVGs of pelicans

I released llm-gemini 0.28 this morning with support for the new model. You can try it out like this:

llm install -U llm-gemini
llm keys set gemini # paste in key
llm -m gemini-3-flash-preview "Generate an SVG of a pelican riding a bicycle"

According to the developer docs the new model supports four different thinking level options: minimal, low, medium, and high. This is different from Gemini 3 Pro, which only supported low and high.

You can run those like this:

llm -m gemini-3-flash-preview --thinking-level minimal "Generate an SVG of a pelican riding a bicycle"

Here are four pelicans, for thinking levels minimal, low, medium, and high:

A minimalist vector illustration of a stylized white bird with a long orange beak and a red cap riding a dark blue bicycle on a single grey ground line against a plain white background. Minimalist illustration: A stylized white bird with a large, wedge-shaped orange beak and a single black dot for an eye rides a red bicycle with black wheels and a yellow pedal against a solid light blue background. A minimalist illustration of a stylized white bird with a large yellow beak riding a red road bicycle in a racing position on a light blue background. Minimalist line-art illustration of a stylized white bird with a large orange beak riding a simple black bicycle with one orange pedal, centered against a light blue circular background.

I built the gallery component with Gemini 3 Flash

The gallery above uses a new Web Component which I built using Gemini 3 Flash to try out its coding abilities. The code on the page looks like this:

<image-gallery width="4">
    <img src="https://static.simonwillison.net/static/2025/gemini-3-flash-preview-thinking-level-minimal-pelican-svg.jpg" alt="A minimalist vector illustration of a stylized white bird with a long orange beak and a red cap riding a dark blue bicycle on a single grey ground line against a plain white background." />
    <img src="https://static.simonwillison.net/static/2025/gemini-3-flash-preview-thinking-level-low-pelican-svg.jpg" alt="Minimalist illustration: A stylized white bird with a large, wedge-shaped orange beak and a single black dot for an eye rides a red bicycle with black wheels and a yellow pedal against a solid light blue background." />
    <img src="https://static.simonwillison.net/static/2025/gemini-3-flash-preview-thinking-level-medium-pelican-svg.jpg" alt="A minimalist illustration of a stylized white bird with a large yellow beak riding a red road bicycle in a racing position on a light blue background." />
    <img src="https://static.simonwillison.net/static/2025/gemini-3-flash-preview-thinking-level-high-pelican-svg.jpg" alt="Minimalist line-art illustration of a stylized white bird with a large orange beak riding a simple black bicycle with one orange pedal, centered against a light blue circular background." />
</image-gallery>

Those alt attributes are all generated by Gemini 3 Flash as well, using this recipe:

llm -m gemini-3-flash-preview --system '
You write alt text for any image pasted in by the user. Alt text is always presented in a
fenced code block to make it easy to copy and paste out. It is always presented on a single
line so it can be used easily in Markdown images. All text on the image (for screenshots etc)
must be exactly included. A short note describing the nature of the image itself should go first.' \
-a https://static.simonwillison.net/static/2025/gemini-3-flash-preview-thinking-level-high-pelican-svg.jpg

You can see the code that powers the image gallery Web Component here on GitHub. I built it by prompting Gemini 3 Flash via LLM like this:

llm -m gemini-3-flash-preview '
Build a Web Component that implements a simple image gallery. Usage is like this:

<image-gallery width="5">
  <img src="image1.jpg" alt="Image 1">
  <img src="image2.jpg" alt="Image 2" data-thumb="image2-thumb.jpg">
  <img src="image3.jpg" alt="Image 3">
</image-gallery>

If an image has a data-thumb= attribute that one is used instead, other images are scaled down. 

The image gallery always takes up 100% of available width. The width="5" attribute means that five images will be shown next to each other in each row. The default is 3. There are gaps between the images. When an image is clicked it opens a modal dialog with the full size image.

Return a complete HTML file with both the implementation of the Web Component several example uses of it. Use https://picsum.photos/300/200 URLs for those example images.'

It took a few follow-up prompts using llm -c:

llm -c 'Use a real modal such that keyboard shortcuts and accessibility features work without extra JS'

llm -c 'Use X for the close icon and make it a bit more subtle'

llm -c 'remove the hover effect entirely'

llm -c 'I want no border on the close icon even when it is focused'

Here's the full transcript, exported using llm logs -cue.

Those five prompts took:

  • 225 input, 3,269 output
  • 2,243 input, 2,908 output
  • 4,319 input, 2,516 output
  • 6,376 input, 2,094 output
  • 8,151 input, 1,806 output

Added together that's 21,314 input and 12,593 output for a grand total of 4.8436 cents.

The guide to migrating from Gemini 2.5 reveals one disappointment:

Image segmentation: Image segmentation capabilities (returning pixel-level masks for objects) are not supported in Gemini 3 Pro or Gemini 3 Flash. For workloads requiring native image segmentation, we recommend continuing to utilize Gemini 2.5 Flash with thinking turned off or Gemini Robotics-ER 1.5.

I wrote about this capability in Gemini 2.5 back in April. I hope they come back in future models - they're a really neat capability that is unique to Gemini.

Tags: google, ai, web-components, generative-ai, llms, llm, gemini, llm-pricing, pelican-riding-a-bicycle, llm-release

GPT-5.2

✇Simon Willison's Weblog
Par : Simon Willison
12 décembre 2025 à 00:58

OpenAI reportedly declared a "code red" on the 1st of December in response to increasingly credible competition from the likes of Google's Gemini 3. It's less than two weeks later and they just announced GPT-5.2, calling it "the most capable model series yet for professional knowledge work".

Key characteristics of GPT-5.2

The new model comes in two variants: GPT-5.2 and GPT-5.2 Pro. There's no Mini variant yet.

GPT-5.2 is available via their UI in both "instant" and "thinking" modes, presumably still corresponding to the API concept of different reasoning effort levels.

The knowledge cut-off date for both variants is now August 31st 2025. This is significant - GPT 5.1 and 5 were both Sep 30, 2024 and GPT-5 mini was May 31, 2024.

Both of the 5.2 models have a 400,000 token context window and 128,000 max output tokens - no different from 5.1 or 5.

Pricing wise 5.2 is a rare increase - it's 1.4x the cost of GPT 5.1, at $1.75/million input and $14/million output. GPT-5.2 Pro is $21.00/million input and a hefty $168.00/million output, putting it up there with their previous most expensive models o1 Pro and GPT-4.5.

So far the main benchmark results we have are self-reported by OpenAI. The most interesting ones are a 70.9% score on their GDPval "Knowledge work tasks" benchmark (GPT-5 got 38.8%) and a 52.9% on ARC-AGI-2 (up from 17.6% for GPT-5.1 Thinking).

The ARC Prize Twitter account provided this interesting note on the efficiency gains for GPT-5.2 Pro

A year ago, we verified a preview of an unreleased version of @OpenAI o3 (High) that scored 88% on ARC-AGI-1 at est. $4.5k/task

Today, we’ve verified a new GPT-5.2 Pro (X-High) SOTA score of 90.5% at $11.64/task

This represents a ~390X efficiency improvement in one year

GPT-5.2 can be accessed in OpenAI's Codex CLI tool like this:

codex -m gpt-5.2

There are three new API models:

  • gpt-5.2 - I think this is what you get if you select "GPT-5.2 Thinking" in ChatGPT but I'm a little confused.
  • gpt-5.2-chat-latest - the model used by ChatGPT for "GPT-5.2 Instant" mode. It's priced the same as GPT-5.2 but has a reduced 128,000 context window with 16,384 max output tokens.
  • gpt-5.2-pro

OpenAI have published a new GPT-5.2 Prompting Guide. An interesting note from that document is that compaction can now be run with a new dedicated server-side API:

For long-running, tool-heavy workflows that exceed the standard context window, GPT-5.2 with Reasoning supports response compaction via the /responses/compact endpoint. Compaction performs a loss-aware compression pass over prior conversation state, returning encrypted, opaque items that preserve task-relevant information while dramatically reducing token footprint. This allows the model to continue reasoning across extended workflows without hitting context limits.

It's better at vision

One note from the announcement that caught my eye:

GPT‑5.2 Thinking is our strongest vision model yet, cutting error rates roughly in half on chart reasoning and software interface understanding.

I had disappointing results from GPT-5 on an OCR task a while ago. I tried it against GPT-5.2 and it did much better:

llm -m gpt-5.2 ocr -a https://static.simonwillison.net/static/2025/ft.jpeg

Here's the result from that, which cost 1,520 input and 1,022 for a total of 1.6968 cents.

Rendering some pelicans

For my classic "Generate an SVG of a pelican riding a bicycle" test:

llm -m gpt-5.2 "Generate an SVG of a pelican riding a bicycle"

Described by GPT-5.2: Cartoon-style illustration: A white, duck-like bird with a small black eye, oversized orange beak (with a pale blue highlight along the lower edge), and a pink neckerchief rides a blue-framed bicycle in side view; the bike has two large black wheels with gray spokes, a blue front fork, visible black crank/pedal area, and thin black handlebar lines, with gray motion streaks and a soft gray shadow under the bike on a light-gray road; background is a pale blue sky with a simple yellow sun at upper left and two rounded white clouds (one near upper center-left and one near upper right).

And for the more advanced alternative test, which tests instruction following in a little more depth:

llm -m gpt-5.2 "Generate an SVG of a California brown pelican riding a bicycle. The bicycle
must have spokes and a correctly shaped bicycle frame. The pelican must have its
characteristic large pouch, and there should be a clear indication of feathers.
The pelican must be clearly pedaling the bicycle. The image should show the full
breeding plumage of the California brown pelican."

Digital illustration on a light gray/white background with a thin horizontal baseline: a stylized California brown pelican in breeding plumage is drawn side-on, leaning forward and pedaling a bicycle; the pelican has a dark brown body with layered wing lines, a pale cream head with a darker brown cap and neck shading, a small black eye, and an oversized long golden-yellow bill extending far past the front wheel; one brown leg reaches down to a pedal while the other is tucked back; the bike is shown in profile with two large spoked wheels (black tires, white rims), a dark frame, crank and chainring near the rear wheel, a black saddle above the rear, and the front fork aligned under the pelican’s head; text at the top reads "California brown pelican (breeding plumage) pedaling a bicycle".

Update 14th December 2025: I used GPT-5.2 running in Codex CLI to port a complex Python library to JavaScript. It ran without interference for nearly four hours and completed a complex task exactly to my specification.

Tags: ai, openai, generative-ai, llms, llm, pelican-riding-a-bicycle, llm-release, gpt-5

Devstral 2

✇Simon Willison's Weblog
Par : Simon Willison
10 décembre 2025 à 00:58

Devstral 2

Two new models from Mistral today: Devstral 2 and Devstral Small 2 - both focused on powering coding agents such as Mistral's newly released Mistral Vibe which I wrote about earlier today.
  • Devstral 2: SOTA open model for code agents with a fraction of the parameters of its competitors and achieving 72.2% on SWE-bench Verified.
  • Up to 7x more cost-efficient than Claude Sonnet at real-world tasks.

Devstral 2 is a 123B model released under a janky license - it's "modified MIT" where the modification is:

You are not authorized to exercise any rights under this license if the global consolidated monthly revenue of your company (or that of your employer) exceeds $20 million (or its equivalent in another currency) for the preceding month. This restriction in (b) applies to the Model and any derivatives, modifications, or combined works based on it, whether provided by Mistral AI or by a third party. [...]

Mistral Small 2 is under a proper Apache 2 license with no weird strings attached. It's a 24B model which is 51.6GB on Hugging Face and should quantize to significantly less.

I tried out the larger model via my llm-mistral plugin like this:

llm install llm-mistral
llm mistral refresh
llm -m mistral/devstral-2512 "Generate an SVG of a pelican riding a bicycle"

Bicycle looks a bit like a cybertruck

For a ~120B model that one is pretty good!

Here's the same prompt with -m mistral/labs-devstral-small-2512 for the API hosted version of Devstral Small 2:

A small white pelican on what looks more like a child's cart.

Again, a decent result given the small parameter size. For comparison, here's what I got for the 24B Mistral Small 3.2 earlier this year.

Tags: ai, generative-ai, llms, llm, mistral, pelican-riding-a-bicycle, llm-release, janky-licenses

❌
❌