Skip to main content

Gemini 3 Pro Deep Think

GoogleCurrentReleased Feb 12, 2026
Overall Score
Est. 86Prov. #8 of 110
Arena Elo
1486
Categories Ranked
8of 8
Price (1M tokens)
$ in / $ out
Speed
N/A
Context
2M
ProprietaryReasoning
Confidence
reasoning

According to BenchLM.ai, Gemini 3 Pro Deep Think ranks #8 out of 110 models on the provisional leaderboard with an overall score of 86/100. It does not yet have enough sourced coverage for BenchLM's verified leaderboard. This places it in the upper tier of AI models, with competitive scores across most benchmark categories.

Gemini 3 Pro Deep Think is a proprietary model with a 2M token context window. It uses explicit chain-of-thought reasoning, which typically improves performance on math and complex reasoning tasks at the cost of higher latency and token usage.

Gemini 3 Pro Deep Think sits inside the Gemini 3 Pro family alongside Gemini 3 Pro. This profile currently has 1 of 152 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.

Its strongest category is Multimodal & Grounded (#2), while its weakest is Instruction Following (#24). This performance profile makes it particularly strong for screenshots, documents, charts, and grounded multimodal workflows.

Ranking Distribution

Category rank across 8 benchmark categories — sorted by best rank

Category Performance

Scores across all benchmark categories (0-100 scale)

Category Breakdown

Agentic

#8
86.0/ 100
Weight: 22%0 benchmarks
Terminal-Bench 2.0BrowseCompOSWorld-VerifiedGAIATAU-benchWebArena

Coding

#22
76.2/ 100
Weight: 20%0 benchmarks
SWE-bench VerifiedLiveCodeBenchSWE-bench ProSWE-RebenchSciCode

Reasoning

#8
88.7/ 100
Weight: 17%1 benchmark
MuSRLongBench v2MRCRv2ARC-AGI-2

Knowledge

#8
88.5/ 100
Weight: 12%0 benchmarks
GPQASuperGPQAMMLU-ProHLEFrontierScienceSimpleQA

Math

#4
96.0/ 100
Weight: 5%0 benchmarks
AIME 2025BRUMO 2025MATH-500FrontierMath

Multilingual

#16
84.6/ 100
Weight: 7%0 benchmarks
MGSMMMLU-ProX

Multimodal

#2
100.0/ 100
Weight: 12%0 benchmarks
MMMU-ProOfficeQA Pro

Inst. Following

#24
82.6/ 100
Weight: 5%0 benchmarks
IFEvalIFBench

Chatbot Arena Performance

Text Overall1486CI: ±3.941,631 votes
Coding1519CI: ±7.28,622 votes
Math1479CI: ±11.62,680 votes
Instruction Following1474CI: ±6.511,259 votes
Creative Writing1485CI: ±8.46,359 votes
Multi-turn1495CI: ±8.06,774 votes
Hard Prompts1504CI: ±5.122,573 votes
Hard Prompts (English)1503CI: ±6.610,776 votes
Longer Query1492CI: ±6.810,592 votes

Benchmark Details

Only benchmark rows with an attached exact-source record are shown here. Source-unverified manual rows and generated rows are hidden from model pages.

Gemini 3 Pro Family

Reasoning

Canonical Entry

Gemini 3 Pro

Frequently Asked Questions

How does Gemini 3 Pro Deep Think perform overall in AI benchmarks?

Gemini 3 Pro Deep Think currently ranks #8 out of 110 models on BenchLM's provisional leaderboard with an overall score of 86 (estimated). It is created by Google and features a 2M context window.

Is Gemini 3 Pro Deep Think good for reasoning and logic?

Gemini 3 Pro Deep Think ranks #8 out of 110 models in reasoning and logic benchmarks with an average score of 88.7. It is among the top performers in this category.

Which sibling models are related to Gemini 3 Pro Deep Think?

Gemini 3 Pro Deep Think belongs to the Gemini 3 Pro family. Related variants on BenchLM include Gemini 3 Pro.

Does Gemini 3 Pro Deep Think have full benchmark coverage on BenchLM?

Not yet. Gemini 3 Pro Deep Think currently has 1 published benchmark scores out of the 152 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.

What is the context window size of Gemini 3 Pro Deep Think?

Gemini 3 Pro Deep Think has a context window of 2M, which determines how much text it can process in a single interaction.

Last updated: April 20, 2026 · Runtime metrics stay blank until BenchLM has a sourced snapshot.

Don't miss the next GPT moment

Which models moved up, what’s new, and what it costs. One email a week, 3-min read.

Free. One email per week.