| Spec | Value |
|---|---|
| Maker | Google DeepMind |
| Context window | 1M tokens |
From Google's launch post and its benchmark table image; the post gives no price, output limit or modalities list, and the Gemini 3 Pro model card PDF could not be read. · source
Estimate from the version's published specs, benchmarks and reviews, 0-5; scores are only comparable within one criteria year. Scored as launched (Gemini 3 Pro preview, 18 Nov 2025). Reasoning: Artificial Analysis Intelligence Index 48 on v4.0 (AA's January 2026 re-baseline, The Decoder 7 Jan 2026; the v3 index at launch gave 73, a scale the anchors do not match, so the v4.0 restatement is used, ~7 weeks after launch). Coding null: Terminal-Bench 4.0 did not exist; Google published only Terminal-Bench 2.0 54.2% (Terminus-2). Context 1M input tokens (Google blog, Simon Willison). Price $2 in / $12 out per 1M up to 200k = $4.50 blended at 3:1 ($4/$18 above 200k). Speed null: Artificial Analysis showed no launch speed. Input: text, image, audio, video. Other figures: LMArena 1501, HLE 37.5%, GPQA Diamond 91.9%.[1]
The best way to describe it is that it's Gemini 2.5 upgraded to match the leading rival models
Gemini 3 Pro is a solid, dependable upgrade with some genuinely impressive highs—especially in frontend user interface work and turning rough prompts into small, working apps.
In Gemini Advanced, I tested code generation and complex multi-step prompts, and the improvement in reasoning was obvious.
We've been testing it and our first impression is that it's a big leap forward for the company's AI tools.
| Trying out Gemini 3 Pro with audio transcription and a new pelican benchmark · Nov 18, 2025 | 0.50 |
|---|---|
| Google's Gemini 3 is here — and it's already beating ChatGPT in key benchmarks · Nov 18, 2025 | 0.70 |
| Vibe Check: Gemini 3 Pro, A Reliable Workhorse With Surprising Flair · Nov 19, 2025 | 0.60 |
| Google Gemini 3 has dropped – here are 6 prompts that show what it can do · Nov 21, 2025 | 0.80 |
Photos from readers, kept apart from the pin's sourced media. Rate what you ate or saw, and on a restaurant pin tag the dishes right on your photo.
The first entry is always the pin's source. Overall confidence is a weighted average of how firmly each reference supports the start and end times used above; a reference counts half as much for every 180 days older than the newest.
Google's post "A new era of intelligence with Gemini 3" is dated 18 November 2025 and says "Today we're releasing Gemini 3", available that day in the Gemini app, AI Mode in Search, the Gemini API and Vertex AI.
The Wikipedia article lists Gemini 3 Pro as released on 18 November 2025 and describes it as a sparse mixture-of-experts, natively multimodal reasoning model.
Something missing or wrong? Say it in your own words: a link that backs this pin up, a different start or end date and why, or a fact it lacks or gets wrong. The AI checks it against this pin's sources, searches for better ones, and adds any page that backs you up. The pin's own sources still count most. A picture that shows something else, or shows it badly, is looked at too, and moved down or replaced.
Epoch AI model database (Gemini 3 Pro); Epoch's confidence in the figures: Unknown. Blank tiles are not published.