Say hello to Gemini 3.7 Flash, generally available today! ⚡️We took your feedback and pushed hard on real-world coding benchmarks and agent autonomy:
- Major jumps on coding and agentic use cases
- DeepSWE (65.3% vs 49.0%) FrontierCode (43.6% vs 34.4%), AutomationBench (30.4% vs 17.0%)
- 50% discount until EOY: $0.75/$3.75 per 1M tokens
- Default model in Managed Agents, @GoogleAIStudio , @antigravity, and @GeminiApp Spark
Throw your hardest coding/agentic tasks at it and let us know what you build: https://t.co/wh09RtN2IK
The Artificial Analysis numbers for Gemini 3.6 Flash are in, and its a mixed bag against its own predecessor.
On overall intelligence, 3.6 Flash and 3.5 Flash are basically tied, both around 50.
On coding its actually a touch worse, 69.2 versus 70.1 for 3.5 Flash.
Where it gains is agentic. 3.6 Flash hits 38.7 to 3.5 Flashs 37.4.
And the real win is cost. It takes $727 to run the full index versus $1,041 for 3.5 Flash. Over $300 cheaper.
So this was a cost and agentic play, not a quality jump. Which lines up with the three.js test earlier, better on paper for agents, weaker on actual rendering.
Gemini 3.6 Flash (HIGH) vs. Gemini 3.5 Flash (HIGH)
Some decent improvements ngl
3.6 flash used only ~6100 tokens to make this whereas 3.5 flash used ~9300 tokens.
3.6 flash is also slightly faster!
Left is 3.6 Flash (NEXUS), right is 3.5 Flash (Synergy).
But I think I like 3.5 flash's look more, but what do you guys think??
Prompt: "create a modern website! Should look amazing! One-code page"
📢GaussianGPT (ECCV'26) Code Release📢
What if 3D scenes worked like language?
Generate full 3D Gaussian scenes - from scratch or from partial - token by token!
🔗https://t.co/LQgrW6sHzt
🌐https://t.co/PseYk8ctIT
Great work by @nicolasvluetzow, @barbara_roessle, @katha_schmid
Grok 4.5 is my default for all my coding work, I freaking love this model. Frontier intelligence that won't break the bank.
Grok 4.5: $2 in / $6 out
GPT 5.6: $5 in / $30 out
Opus 4.8: $5 in / $25 out