DeepSeek-V4-Pro vs Qwen3.8 Max Preview
Wins 0 of 7 areas
—
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Math · Long documents · Following instructions
Qwen3.8 Max Preview is the stronger all-rounder.DeepSeek-V4-Pro is cheaper.
Scores updated · 38 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.
- CodingWriting and fixing software06Qwen3.8 Max Preview6 of 6 tests
- AgentsCarrying out multi-step tasks on its own05Qwen3.8 Max Preview5 of 5 tests
- ReasoningHard problems that need careful thinking04Qwen3.8 Max Preview4 of 4 tests
- MathCompetition and research-level math02Qwen3.8 Max Preview2 of 3 tests · 1 tie
- Following instructionsDoing exactly what it is asked02Qwen3.8 Max Preview2 of 2 tests
- FactsGetting facts right instead of making them up12Qwen3.8 Max Preview2 of 3 tests
- Long documentsFinding answers in very long texts01Qwen3.8 Max Preview1 of 1 test
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-Pro costs 84% 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 Qwen3.8 Max Preview pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+65.3points ahead
- Extremely hard research-level math problemsFrontierMath Tier 4+43.9points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+29.4points ahead
Where DeepSeek-V4-Pro pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+11.3points ahead
Every test, side by side
All 38 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.8 Max Preview
- Terminal-Bench 2.1Qwen3.8 Max Preview by 24.86488.8+24.8
- LiveBench · Agentic CodingQwen3.8 Max Preview by 22.142.664.7+22.1
- LMArena · WebDevQwen3.8 Max Preview by 208 rating points14641672+208 rating
- SWE-bench ProQwen3.8 Max Preview by 12.355.467.7+12.3
- LiveBench · CodingQwen3.8 Max Preview by 2.97072.9+2.9
- SciCodeQwen3.8 Max Preview by 1.350.852.1+1.3
AgentsQwen3.8 Max Preview
- GDPValQwen3.8 Max Preview by 25.532.858.3+25.5
- Terminal-Bench 4.0Qwen3.8 Max Preview by 24.314.638.9+24.3
- AA ApexAgentsQwen3.8 Max Preview by 18.124.342.4+18.1
- τ-Bench V3 · BankingQwen3.8 Max Preview by 17.730.147.8+17.7
- AA IT-Bench SREQwen3.8 Max Preview by 238.340.3+2
ReasoningQwen3.8 Max Preview
- Humanity's Last ExamQwen3.8 Max Preview by 5.637.543.1+5.6
- LiveBench · ReasoningQwen3.8 Max Preview by 5.582.788.2+5.5
- CritPtQwen3.8 Max Preview by 4.812.917.7+4.8
- GPQA DiamondQwen3.8 Max Preview by 488.892.8+4
FactsQwen3.8 Max Preview
- AA-Omniscience · Non-hallucinationQwen3.8 Max Preview by 65.35.971.2+65.3
- AA-Omniscience · AccuracyDeepSeek-V4-Pro by 11.34331.7+11.3
- SimpleQA VerifiedQwen3.8 Max Preview by 1.146.247.3+1.1
MathQwen3.8 Max Preview
- FrontierMath Tier 4Qwen3.8 Max Preview by 43.92.446.3+43.9
- FrontierMath Tiers 1-3 (v2)Qwen3.8 Max Preview by 29.445.374.7+29.4
- LiveBench · Mathematicstie90.791.3tie
Long documentsQwen3.8 Max Preview
- AA-LCRQwen3.8 Max Preview by 5.674.780.3+5.6
Following instructionsQwen3.8 Max Preview
- LiveBench · Instruction FollowingQwen3.8 Max Preview by 11.762.474.1+11.7
- IFBenchQwen3.8 Max Preview by 6.376.582.8+6.3
Other results14 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.
- DeepSWEQwen3.8 Max Preview by 43.812.856.6+43.8
- GDPval-AA v2Qwen3.8 Max Preview by 433 rating points13061739+433 rating
- Agents' Last ExamQwen3.8 Max Preview by 26.725.752.4+26.7
- AA-OmniscienceQwen3.8 Max Preview by 22.7-10.712+22.7
- AA Agentic IndexQwen3.8 Max Preview by 21.927.749.6+21.9
- ToolathlonQwen3.8 Max Preview by 20.751.872.5+20.7
- NL2RepoQwen3.8 Max Preview by 17.438.555.9+17.4
- Toolathlon VerifiedQwen3.8 Max Preview by 16.655.972.5+16.6
- AA IntelligenceQwen3.8 Max Preview by 1530.445.4+15
- Artificial Analysis Coding IndexQwen3.8 Max Preview by 12.459.471.8+12.4
- CyberGymDeepSeek-V4-Pro by 4.883.378.5+4.8
- livebench_data_analysisQwen3.8 Max Preview by 3.974.578.4+3.9
- HLE (with tools)DeepSeek-V4-Pro by 3.86056.2+3.8
- livebench_languageQwen3.8 Max Preview by 1.678.179.7+1.6
Questions people ask
Which is better, DeepSeek-V4-Pro or Qwen3.8 Max Preview?
Qwen3.8 Max Preview wins all seven areas where both have results: coding, agents, reasoning, facts, math, long documents and following instructions. DeepSeek-V4-Pro wins none, but costs 84% less.
Which is better for coding?
Qwen3.8 Max Preview. It wins 6 of the 6 coding tests both models report; DeepSeek-V4-Pro wins none.
Which is cheaper?
DeepSeek-V4-Pro costs $0.43 per million input tokens and $0.87 per million output tokens; Qwen3.8 Max Preview costs $2.00 and $6.00. That makes DeepSeek-V4-Pro about 84% cheaper for the same work.
How do you compare the two?
We use the 38 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 14 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.