DeepSeek-V4-Flash vs Qwen3.6 Plus
Wins 5 of 7 areas
Agents · Reasoning · Math · Long documents · Following instructions
Wins 1 of 7 areas
Facts
DeepSeek-V4-Flash is the stronger all-rounder.Qwen3.6 Plus is better at facts.
Scores updated · 40 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.
- MathCompetition and research-level math40DeepSeek-V4-Flash4 of 5 tests · 1 tie
- ReasoningHard problems that need careful thinking40DeepSeek-V4-Flash4 of 4 tests
- Following instructionsDoing exactly what it is asked20DeepSeek-V4-Flash2 of 2 tests
- AgentsCarrying out multi-step tasks on its own21DeepSeek-V4-Flash2 of 3 tests
- Long documentsFinding answers in very long texts10DeepSeek-V4-Flash1 of 1 test
- FactsGetting facts right instead of making them up12Qwen3.6 Plus2 of 3 tests
- CodingWriting and fixing software44Even4 each · 2 ties
Long documents rests on a single test.
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 50% 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
- Real work tasks from 44 professionsGDPVal+22.2points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+19.2points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+18.6points ahead
Where Qwen3.6 Plus pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+57.1points ahead
- Short factual questions, answered correctlySimpleQA Verified+10points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+8.3points ahead
Every test, side by side
All 40 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingEven
- Terminal-Bench 2.1DeepSeek-V4-Flash by 17.378.761.4+17.3
- SciCodeDeepSeek-V4-Flash by 9.650.340.7+9.6
- Terminal-Bench HardQwen3.6 Plus by 8.335.643.9+8.3
- LiveBench · Agentic CodingDeepSeek-V4-Flash by 5.446.841.4+5.4
- SWE-bench ProQwen3.6 Plus by 452.656.6+4
- LiveCodeBench v6DeepSeek-V4-Flash by 3.590.687.1+3.5
- LiveBench · CodingQwen3.6 Plus by 3.27578.2+3.2
- LMArena · WebDevQwen3.6 Plus by 31 rating points14301461+31 rating
- SWE-bench Multilingualtie73.373.8tie
- SWE-bench Verifiedtie7978.8tie
AgentsDeepSeek-V4-Flash
- GDPValDeepSeek-V4-Flash by 22.246.924.7+22.2
- τ-Bench V3 · BankingDeepSeek-V4-Flash by 18.639.420.8+18.6
- MCP AtlasQwen3.6 Plus by 5.16974.1+5.1
ReasoningDeepSeek-V4-Flash
- CritPtDeepSeek-V4-Flash by 13.716.62.9+13.7
- Humanity's Last ExamDeepSeek-V4-Flash by 10.838.627.8+10.8
- LiveBench · ReasoningDeepSeek-V4-Flash by 10.886.675.8+10.8
- GPQA DiamondDeepSeek-V4-Flash by 2.690.888.2+2.6
FactsQwen3.6 Plus
- AA-Omniscience · Non-hallucinationQwen3.6 Plus by 57.18.365.4+57.1
- AA-Omniscience · AccuracyDeepSeek-V4-Flash by 1440.426.4+14
- SimpleQA VerifiedQwen3.6 Plus by 1034.144.1+10
MathDeepSeek-V4-Flash
- FrontierMath Tiers 1-3 (v2)DeepSeek-V4-Flash by 19.257.538.3+19.2
- HMMT Feb. 2026DeepSeek-V4-Flash by 794.887.8+7
- IMOAnswerBenchDeepSeek-V4-Flash by 4.688.483.8+4.6
- LiveBench · MathematicsDeepSeek-V4-Flash by 3.186.883.7+3.1
- AIME 2026tie95.895.3tie
Long documentsDeepSeek-V4-Flash
- AA-LCRDeepSeek-V4-Flash by 1.479.778.3+1.4
Following instructionsDeepSeek-V4-Flash
- LiveBench · Instruction FollowingDeepSeek-V4-Flash by 7.265.558.3+7.2
- IFBenchDeepSeek-V4-Flash by 479.275.2+4
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.
- ToolathlonDeepSeek-V4-Flash by 30.570.339.8+30.5
- NL2RepoDeepSeek-V4-Flash by 16.354.237.9+16.3
- AA-OmniscienceQwen3.6 Plus by 15.2-14.30.9+15.2
- Artificial Analysis Coding IndexDeepSeek-V4-Flash by 14.669.154.5+14.6
- AA Agentic IndexDeepSeek-V4-Flash by 12.741.729+12.7
- livebench_data_analysisDeepSeek-V4-Flash by 9.479.369.9+9.4
- AA IntelligenceDeepSeek-V4-Flash by 7.334.327+7.3
- HLE (with tools)Qwen3.6 Plus by 5.545.150.6+5.5
- Terminal-Bench 2.0Qwen3.6 Plus by 4.756.961.6+4.7
- livebench_languageDeepSeek-V4-Flash by 4.279.275+4.2
- τ²-Bench Telecom (AA run)Qwen3.6 Plus by 2.79597.7+2.7
- MMLU-ProQwen3.6 Plus by 2.386.288.5+2.3
Questions people ask
Which is better, DeepSeek-V4-Flash or Qwen3.6 Plus?
DeepSeek-V4-Flash wins five of the seven areas where both have results: agents, reasoning, math, long documents and following instructions. Qwen3.6 Plus wins facts. They are level on coding.
Which is better for coding?
Neither. They win 4 coding tests each of the 10 both models report, and 2 are ties.
Which is cheaper?
DeepSeek-V4-Flash costs $0.44 per million input tokens and $1.32 per million output tokens; Qwen3.6 Plus costs $0.50 and $3.00. That makes DeepSeek-V4-Flash about 50% cheaper for the same work.
How do you compare the two?
We use the 40 benchmark tests both models have published scores on. The verdict counts the 28 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.