DeepSeek-V3.2-Speciale vs Gemini 3 Pro
Wins 0 of 6 areas
—
Wins 5 of 6 areas
Coding · Agents · Reasoning · Long documents · Following instructions
Gemini 3 Pro is the stronger all-rounder.DeepSeek-V3.2-Speciale is cheaper.
Scores updated · 24 tests both models report · How we compare
Where each one wins
Tests won in each of the six areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- ReasoningHard problems that need careful thinking04Gemini 3 Pro4 of 4 tests
- CodingWriting and fixing software03Gemini 3 Pro3 of 3 tests
- AgentsCarrying out multi-step tasks on its own01Gemini 3 Pro1 of 1 test
- Long documentsFinding answers in very long texts01Gemini 3 Pro1 of 1 test
- Following instructionsDoing exactly what it is asked01Gemini 3 Pro1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
Agents, long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
DeepSeek-V3.2-Speciale costs 95% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Gemini 3 Pro pulls ahead
- Common-sense trick questionsSimpleBench+23.8points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+18points ahead
- Code for real scientific research problemsSciCode+12.1points ahead
Where DeepSeek-V3.2-Speciale pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+2.3points ahead
Every test, side by side
All 24 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingGemini 3 Pro
- SciCodeGemini 3 Pro by 12.14456.1+12.1
- Terminal-Bench HardGemini 3 Pro by 6.934.841.7+6.9
- SWE-bench VerifiedGemini 3 Pro by 3.173.176.2+3.1
ReasoningGemini 3 Pro
- SimpleBenchGemini 3 Pro by 23.852.676.4+23.8
- Humanity's Last ExamGemini 3 Pro by 1128.739.7+11
- GPQA DiamondGemini 3 Pro by 3.787.190.8+3.7
- CritPtGemini 3 Pro by 1.77.49.1+1.7
FactsEven
- AA-Omniscience · AccuracyGemini 3 Pro by 1837.855.8+18
- AA-Omniscience · Non-hallucinationDeepSeek-V3.2-Speciale by 2.310.88.5+2.3
Following instructionsGemini 3 Pro
- IFBenchGemini 3 Pro by 6.563.970.4+6.5
Other results12 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- τ²-Bench Telecom (AA run)Gemini 3 Pro by 87.1087.1+87.1
- AA Agentic IndexGemini 3 Pro by 52052+52
- AA-OmniscienceGemini 3 Pro by 33-17.715.3+33
- putnam_2025Gemini 3 Pro by 16.659.275.8+16.6
- AA IntelligenceGemini 3 Pro by 13.514.528+13.5
- Artificial Analysis Coding IndexGemini 3 Pro by 8.637.946.5+8.6
- Terminal-Bench 2.0Gemini 3 Pro by 7.846.454.2+7.8
- t2-benchGemini 3 Pro by 5.180.385.4+5.1
- AIME 2025Gemini 3 Pro by 496100+4
- ToolathlonGemini 3 Pro by 1.235.236.4+1.2
- HMMT Feb. 2025tie97.597.5tie
- HMMT Nov. 2025tie93.393.3tie
Questions people ask
Which is better, DeepSeek-V3.2-Speciale or Gemini 3 Pro?
Gemini 3 Pro wins five of the six areas where both have results: coding, agents, reasoning, long documents and following instructions. DeepSeek-V3.2-Speciale wins none, but costs 95% less. They are level on facts.
Which is better for coding?
Gemini 3 Pro. It wins 3 of the 3 coding tests both models report; DeepSeek-V3.2-Speciale wins none.
Which is cheaper?
DeepSeek-V3.2-Speciale costs $0.29 per million input tokens and $0.43 per million output tokens; Gemini 3 Pro costs $2.00 and $12.00. That makes DeepSeek-V3.2-Speciale about 95% cheaper for the same work.
How do you compare the two?
We use the 24 benchmark tests both models have published scores on. The verdict counts the 12 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 12 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.