AI Model Release Tracker - Timeline of Major AI Models from 2022-2026

DeepSeek-V4-Pro vs Gemini 3.5 Flash

DeepSeek
DeepSeek-V4-Pro
Google
Gemini 3.5 Flash
Overview
CompanyDeepSeekGoogle
Release dateApr 24 2026May 19 2026
AccessOpen WeightProprietary
Benchmarks
Nonsense detection
BullshitBench v2
14%
20%Best
Agentic coding
SWE-Bench Pro
55.1%
Agentic coding
CursorBench v3.1
49.8%
Agentic terminal coding
Terminal-Bench 2.1
76.2%
Multi-step tool use
MCP Atlas
83.6%
General tool use
Toolathlon
56.5%
Multidisciplinary reasoning
Humanity's Last Exam · no tools
40.2%
Abstract reasoning
ARC-AGI-2
72.1%
Science
GPQA Diamond
90.1%
Agentic computer use
OSWorld-Verified
78.4%
Agentic financial analysis
Finance Agent v2
57.9%
Knowledge work
GDPval-AA
1656
Chart reasoning
CharXiv Reasoning
84.2%
Multimodal reasoning
MMMU-Pro
83.6%
Spatial reasoning
Blueprint-Bench 2
33.6%
Long context
MRCR v2 (8-needle) · 128k average
77.3%
Long context
MRCR v2 (8-needle) · 1M pointwise
26.6%
Timeline
Release gapDeepSeek-V4-Pro shipped 25 days before Gemini 3.5 Flash

Which is better: DeepSeek-V4-Pro or Gemini 3.5 Flash?

Gemini 3.5 Flash leads DeepSeek-V4-Pro on 1 of the 1 benchmark they both report (BullshitBench v2). DeepSeek-V4-Pro shipped 25 days before Gemini 3.5 Flash, so benchmark comparisons should account for the intervening progress.

DeepSeek-V4-Pro is open weight, while Gemini 3.5 Flash is proprietary.

On BullshitBench v2, Gemini 3.5 Flash leads at 20% vs DeepSeek-V4-Pro at 14%.

Frequently asked questions

DeepSeek-V4-Pro was released by DeepSeek on Apr 24 2026.

Gemini 3.5 Flash was released by Google on May 19 2026.

DeepSeek-V4-Pro is an open weight model released by DeepSeek. Gemini 3.5 Flash is a proprietary model released by Google.

Other comparisons