DeepSeek-V4-Flash vs Gemini 3.5 Flash Lite
Wins 5 of 7 areas
Coding · Agents · Reasoning · Math · Long documents
Wins 1 of 7 areas
Following instructions
DeepSeek-V4-Flash is the stronger all-rounder.Gemini 3.5 Flash Lite is better at following instructions.
Scores updated · 36 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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 thinking50DeepSeek-V4-Flash5 of 5 tests
- MathCompetition and research-level math50DeepSeek-V4-Flash5 of 5 tests
- AgentsCarrying out multi-step tasks on its own30DeepSeek-V4-Flash3 of 3 tests
- CodingWriting and fixing software43DeepSeek-V4-Flash4 of 7 tests
- Long documentsFinding answers in very long texts10DeepSeek-V4-Flash1 of 1 test
- Following instructionsDoing exactly what it is asked01Gemini 3.5 Flash Lite1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
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-V4-Flash costs 37% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where DeepSeek-V4-Flash pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+51.1points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+31.5points ahead
- Harvard-MIT high-school math contest problemsHMMT Feb. 2026+31.2points ahead
Where Gemini 3.5 Flash Lite pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+57.3points ahead
- Follows detailed instructions exactlyLiveBench · Instruction Following+1.7points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+1.6points ahead
Every test, side by side
All 36 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-V4-Flash
- Terminal-Bench 2.1DeepSeek-V4-Flash by 25.178.753.6+25.1
- SciCodeDeepSeek-V4-Flash by 950.341.3+9
- SWE-bench VerifiedDeepSeek-V4-Flash by 47975+4
- SWE-bench ProGemini 3.5 Flash Lite by 1.652.654.2+1.6
- LiveBench · Agentic CodingDeepSeek-V4-Flash by 1.546.845.3+1.5
- LiveBench · CodingGemini 3.5 Flash Lite by 1.17576.1+1.1
- LMArena · WebDevGemini 3.5 Flash Lite by 10 rating points14301440+10 rating
AgentsDeepSeek-V4-Flash
- GDPValDeepSeek-V4-Flash by 22.646.924.3+22.6
- τ-Bench V3 · BankingDeepSeek-V4-Flash by 21.939.417.5+21.9
- Terminal-Bench 4.0DeepSeek-V4-Flash by 11.112.11+11.1
ReasoningDeepSeek-V4-Flash
- ARC-AGI-2DeepSeek-V4-Flash by 51.161.410.3+51.1
- LiveBench · ReasoningDeepSeek-V4-Flash by 26.486.660.2+26.4
- Humanity's Last ExamDeepSeek-V4-Flash by 19.838.618.8+19.8
- CritPtDeepSeek-V4-Flash by 16.616.60+16.6
- GPQA DiamondDeepSeek-V4-Flash by 790.883.8+7
FactsEven
- AA-Omniscience · Non-hallucinationGemini 3.5 Flash Lite by 57.38.365.6+57.3
- AA-Omniscience · AccuracyDeepSeek-V4-Flash by 10.940.429.5+10.9
MathDeepSeek-V4-Flash
- FrontierMath Tiers 1-3 (v2)DeepSeek-V4-Flash by 31.557.526+31.5
- HMMT Feb. 2026DeepSeek-V4-Flash by 31.294.863.6+31.2
- FrontierMath Tier 4DeepSeek-V4-Flash by 24.424.40+24.4
- AIME 2026DeepSeek-V4-Flash by 13.695.882.2+13.6
- LiveBench · MathematicsDeepSeek-V4-Flash by 13.186.873.7+13.1
Following instructionsGemini 3.5 Flash Lite
- LiveBench · Instruction FollowingGemini 3.5 Flash Lite by 1.765.567.2+1.7
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.
- ARC-AGI-1DeepSeek-V4-Flash by 35.58953.5+35.5
- livebench_data_analysisDeepSeek-V4-Flash by 2679.353.3+26
- AA Agentic IndexDeepSeek-V4-Flash by 25.841.715.9+25.8
- Artificial Analysis Coding IndexDeepSeek-V4-Flash by 19.869.149.3+19.8
- AA-OmniscienceGemini 3.5 Flash Lite by 19.5-14.35.2+19.5
- Toolathlon VerifiedDeepSeek-V4-Flash by 13.270.357.1+13.2
- AA IntelligenceDeepSeek-V4-Flash by 12.134.322.2+12.1
- livebench_languageDeepSeek-V4-Flash by 7.479.271.8+7.4
- τ³-Bench BankingDeepSeek-V4-Flash by 6.422.916.5+6.4
- GDPval-AA v2DeepSeek-V4-Flash by 49 rating points11891140+49 rating
- HLE (with tools)DeepSeek-V4-Flash by 2.645.142.5+2.6
- SWEBench Pro PublicGemini 3.5 Flash Lite by 1.652.654.2+1.6
Questions people ask
Which is better, DeepSeek-V4-Flash or Gemini 3.5 Flash Lite?
DeepSeek-V4-Flash wins five of the seven areas where both have results: coding, agents, reasoning, math and long documents. Gemini 3.5 Flash Lite wins following instructions. They are level on facts.
Which is better for coding?
DeepSeek-V4-Flash. It wins 4 of the 7 coding tests both models report; Gemini 3.5 Flash Lite wins 3.
Which is cheaper?
DeepSeek-V4-Flash costs $0.44 per million input tokens and $1.32 per million output tokens; Gemini 3.5 Flash Lite costs $0.30 and $2.50. That makes DeepSeek-V4-Flash about 37% cheaper for the same work.
How do you compare the two?
We use the 36 benchmark tests both models have published scores on. The verdict counts the 24 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.