← All Comparisons

Gemini 3.1 Pro vs DeepSeek R1

A detailed comparison of Gemini 3.1 Pro (Google) and DeepSeek R1 (DeepSeek) across pricing, performance, and features.

Pricing Comparison

MetricGemini 3.1 ProDeepSeek R1Difference
Input / 1M tokens$2.00$0.55-73%
Output / 1M tokens$12.00$2.19-82%
Context window1M128K
Max output64K64K

Benchmark Comparison

BenchmarkGemini 3.1 ProDeepSeek R1
MMLU-Pro91%84%
HumanEval95%92%
GPQA94.3%71.5%

Capabilities

CapabilityGemini 3.1 ProDeepSeek R1
audio
code
reasoning
text
tool-use
vision

Gemini 3.1 Pro Strengths

  • #1 on 12 of 18 tracked benchmarks
  • 94.3% GPQA Diamond — highest of any model
  • Same price as Gemini 3 Pro (free upgrade)
  • 1M context with configurable thinking levels

Gemini 3.1 Pro Weaknesses

  • Still in preview
  • Context-tiered pricing ($4/$18 above 200K)

DeepSeek R1 Strengths

  • Cheapest reasoning model available
  • Strong math and science performance
  • Open-source with off-peak discounts

DeepSeek R1 Weaknesses

  • Slower than non-reasoning models
  • No vision or tool-use
  • China-based — availability concerns

Quick Verdict

Best value: DeepSeek R1 is the more affordable option at $0.55/$2.19 per 1M tokens.

Higher benchmarks: Gemini 3.1 Pro scores higher on average across available benchmarks (93.4% avg).

Larger context: Gemini 3.1 Pro supports 1M tokens.

Choose DeepSeek R1 if cost matters most. Choose Gemini 3.1 Pro if you need the best possible quality for complex tasks.

More Comparisons