Compare · ModelsLive · 2 picked · head to head

Gemini 3.1 Flash Lite vs o4 Mini

Side by side · benchmarks, pricing, and signals you can act on.

Winner summary

o4 Mini wins 5 of 9 shared benchmarks. Leads in general · math · coding.

Category leads
knowledge·Gemini 3.1 Flash Litegeneral·o4 Minimath·o4 Minicoding·o4 Mini

Head to head · from the data · data as of 2026-11-22

  • Benchmarks · Gemini 3.1 Flash Lite leads on 4 of 9 shared benchmarks (Chess Puzzles, GeoBench, GPQA diamond, Lmca), o4 Mini on 5 (Dtbench, FrontierMath-Tiers-1-3-v2-Private, HLE, OTIS Mock AIME 2024-2025, 1 more)
  • Price · Gemini 3.1 Flash Lite input $0.25/M vs o4 Mini $1.10/M, Gemini 3.1 Flash Lite is 77% cheaper (OpenRouter list price, as of 2026-11-22)
  • Cheapest provider · Gemini 3.1 Flash Lite at Google and 1 other provider $0.13/M input (of 6 offers) as of 2026-11-22
  • Context · Gemini 3.1 Flash Lite 1.0M tokens vs o4 Mini 200K tokens
  • Released · Gemini 3.1 Flash Lite May 7, 2026 vs o4 Mini Apr 16, 2025 · 13 months apart
  • Price history · Gemini 3.1 Flash Lite OpenRouter input list price had no move of 5% or more in the 90 days to 2026-11-22 · o4 Mini OpenRouter input list price had no move of 5% or more in the 90 days to 2026-11-22

Sources · scores from public leaderboards, list prices and provider offers from OpenRouter, price moves from the BenchGecko changelog (5% or more, held in the next daily check), Gecko tests run by BenchGecko.

Hype vs Reality
Gemini 3.1 Flash Lite
#146 by perf·#11 by attention
UNDERRATED
o4 Mini
#149 by perf·no signal
QUIET
Best value
3.2x better value than o4 Mini
Gemini 3.1 Flash Lite
55.8 pts/$
$0.88/M
o4 Mini
17.7 pts/$
$2.75/M
Vendor risk
Google DeepMind logo
Google DeepMind
$3.26T·Tier 1
Low risk
OpenAI logo
OpenAI
$840.0B·Tier 1
Medium risk
Head to head
Gemini 3.1 Flash Liteo4 Mini
Chess Puzzles
Gemini 3.1 Flash Lite leads by +3.2
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Gemini 3.1 Flash Lite
21.1
o4 Mini
17.9
Dtbench
o4 Mini leads by +1.3
Gemini 3.1 Flash Lite
61.3
o4 Mini
62.7
FrontierMath-Tiers-1-3-v2-Private
o4 Mini leads by +8.9
Gemini 3.1 Flash Lite
29.2
o4 Mini
38.0
GeoBench
Gemini 3.1 Flash Lite leads by +19.0
GeoBench · tests geographic knowledge and spatial reasoning across countries, landmarks, coordinates, and geopolitical understanding.
Gemini 3.1 Flash Lite
83.0
o4 Mini
64.0
GPQA diamond
Gemini 3.1 Flash Lite leads by +3.1
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Gemini 3.1 Flash Lite
78.9
o4 Mini
75.8
HLE
o4 Mini leads by +9.9
HLE (Humanity's Last Exam) · a reasoning benchmark designed to be the hardest public evaluation of AI. Questions span mathematics, physics, philosophy, and logic · curated to be at or beyond the frontier of human expert capability. Tested with and without tool augmentation.
Gemini 3.1 Flash Lite
4.0
o4 Mini
13.9
Lmca
Gemini 3.1 Flash Lite leads by +10.0
Gemini 3.1 Flash Lite
41.2
o4 Mini
31.2
OTIS Mock AIME 2024-2025
o4 Mini leads by +1.7
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Gemini 3.1 Flash Lite
80.0
o4 Mini
81.7
WeirdML
o4 Mini leads by +0.4
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Gemini 3.1 Flash Lite
52.2
o4 Mini
52.6
Full benchmark table
BenchmarkGemini 3.1 Flash Liteo4 Mini
Chess Puzzles
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
21.117.9
Dtbench
61.362.7
FrontierMath-Tiers-1-3-v2-Private
29.238.0
GeoBench
GeoBench · tests geographic knowledge and spatial reasoning across countries, landmarks, coordinates, and geopolitical understanding.
83.064.0
GPQA diamond
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
78.975.8
HLE
HLE (Humanity's Last Exam) · a reasoning benchmark designed to be the hardest public evaluation of AI. Questions span mathematics, physics, philosophy, and logic · curated to be at or beyond the frontier of human expert capability. Tested with and without tool augmentation.
4.013.9
Lmca
41.231.2
OTIS Mock AIME 2024-2025
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
80.081.7
WeirdML
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
52.252.6
Pricing · per 1M tokens · projected $/mo at 10M tokens
ModelInputOutputContextProjected $/mo
Google DeepMind logoGemini 3.1 Flash Lite$0.25$1.501.0M tokens (~524 books)$5.63
OpenAI logoo4 Mini$1.10$4.40200K tokens (~100 books)$19.25
Gemini 3.1 Flash Lite leads on 4 of 9 shared benchmarks (Chess Puzzles, GeoBench, GPQA diamond, Lmca), o4 Mini on 5 (Dtbench, FrontierMath-Tiers-1-3-v2-Private, HLE, OTIS Mock AIME 2024-2025, 1 more)
People also compared