Qwen3.8 27B vs Qwen3.8 Max Preview
Wins 0 of 8 areas
—
Wins 7 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math · Following instructions
Qwen3.8 Max Preview is the stronger all-rounder.Qwen3.8 27B is cheaper.
Both rank among the ten best models we track in following instructions.
Scores updated · 42 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. 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 software15Qwen3.8 Max Preview5 of 6 tests
- ReasoningHard problems that need careful thinking04Qwen3.8 Max Preview4 of 4 tests
- AgentsCarrying out multi-step tasks on its own03Qwen3.8 Max Preview3 of 4 tests · 1 tie
- FactsGetting facts right instead of making them up02Qwen3.8 Max Preview2 of 2 tests
- Images and chartsUnderstanding pictures, charts and video02Qwen3.8 Max Preview2 of 2 tests
- Following instructionsDoing exactly what it is asked02Qwen3.8 Max Preview2 of 2 tests
- MathCompetition and research-level math01Qwen3.8 Max Preview1 of 1 test
- Long documentsFinding answers in very long texts11Even1 each
Math 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.
Qwen3.8 27B costs 56% 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
- Complex command-line tasks across many fieldsTerminal-Bench 4.0+33.3points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+16.1points ahead
- Questions about very long textsLongBench v2+14points ahead
Where Qwen3.8 27B pulls ahead
- Programming problems, refreshed regularlyLiveBench · Coding+2.8points ahead
- Reasons across sets of long documentsAA-LCR+1.7points ahead
Every test, side by side
All 42 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 979.888.8+9
- LMArena · WebDevQwen3.8 Max Preview by 80 rating points15901670+80 rating
- SWE-bench ProQwen3.8 Max Preview by 661.767.7+6
- SciCodeQwen3.8 Max Preview by 5.546.652.1+5.5
- LiveBench · Agentic CodingQwen3.8 Max Preview by 3.361.464.7+3.3
- LiveBench · CodingQwen3.8 27B by 2.875.772.9+2.8
AgentsQwen3.8 Max Preview
- Terminal-Bench 4.0Qwen3.8 Max Preview by 33.35.638.9+33.3
- GDPValQwen3.8 Max Preview by 12.34658.3+12.3
- OSWorld-VerifiedQwen3.8 Max Preview by 1.884.386.1+1.8
- τ-Bench V3 · Bankingtie4847.8tie
ReasoningQwen3.8 Max Preview
- CritPtQwen3.8 Max Preview by 12.35.417.7+12.3
- Humanity's Last ExamQwen3.8 Max Preview by 9.233.943.1+9.2
- LiveBench · ReasoningQwen3.8 Max Preview by 8.28088.2+8.2
- GPQA DiamondQwen3.8 Max Preview by 2.390.592.8+2.3
FactsQwen3.8 Max Preview
- AA-Omniscience · AccuracyQwen3.8 Max Preview by 16.115.631.7+16.1
- AA-Omniscience · Non-hallucinationQwen3.8 Max Preview by 1.569.771.2+1.5
Images and chartsQwen3.8 Max Preview
- MMMU-ProQwen3.8 Max Preview by 6.576.382.8+6.5
- LMArena · VisionQwen3.8 Max Preview by 36 rating points12711307+36 rating
MathQwen3.8 Max Preview
- LiveBench · MathematicsQwen3.8 Max Preview by 5.186.291.3+5.1
Long documentsEven
- LongBench v2Qwen3.8 Max Preview by 1452.366.3+14
- AA-LCRQwen3.8 27B by 1.78280.3+1.7
Following instructionsQwen3.8 Max Preview
- IFBenchQwen3.8 Max Preview by 3.379.582.8+3.3
- LiveBench · Instruction FollowingQwen3.8 Max Preview by 1.472.774.1+1.4
Other results19 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.
- CritPt v2 HardQwen3.8 Max Preview by 25.5429.5+25.5
- AA-OmniscienceQwen3.8 Max Preview by 22-1012+22
- JobBenchQwen3.8 Max Preview by 2033.453.4+20
- DeepSWE 1.1Qwen3.8 Max Preview by 14.442.256.6+14.4
- NL2RepoQwen3.8 Max Preview by 13.642.355.9+13.6
- ERQAQwen3.8 Max Preview by 12.365.577.8+12.3
- Terminal-Bench-Science 0.1Qwen3.8 Max Preview by 11.9011.9+11.9
- AA IntelligenceQwen3.8 Max Preview by 11.733.745.4+11.7
- Agents' Last ExamQwen3.8 Max Preview by 9.542.952.4+9.5
- LVBenchQwen3.8 Max Preview by 9.472.481.8+9.4
- HLE-DiamondQwen3.8 Max Preview by 8.715.324+8.7
- GDP (Surge AI)Qwen3.8 Max Preview by 6.416.422.8+6.4
- Toolathlon VerifiedQwen3.8 Max Preview by 5.467.172.5+5.4
- livebench_languageQwen3.8 Max Preview by 5.374.379.7+5.3
- Artificial Analysis Coding IndexQwen3.8 Max Preview by 3.768.171.8+3.7
- AndroidWorldQwen3.8 Max Preview by 3.481.985.3+3.4
- AA Agentic IndexQwen3.8 Max Preview by 3.146.549.6+3.1
- RealWorldQAQwen3.8 Max Preview by 2.185.988+2.1
- livebench_data_analysisQwen3.8 Max Preview by 1.876.678.4+1.8
Questions people ask
Which is better, Qwen3.8 27B or Qwen3.8 Max Preview?
Qwen3.8 Max Preview wins seven of the eight areas we test: coding, agents, reasoning, facts, images and charts, math and following instructions. Qwen3.8 27B wins none, but costs 56% less. They are level on long documents.
Which is better for coding?
Qwen3.8 Max Preview. It wins 5 of the 6 coding tests both models report; Qwen3.8 27B wins 1.
Which is cheaper?
Qwen3.8 27B costs $0.50 per million input tokens and $3.00 per million output tokens; Qwen3.8 Max Preview costs $2.00 and $6.00. That makes Qwen3.8 27B about 56% cheaper for the same work.
How do you compare the two?
We use the 42 benchmark tests both models have published scores on. The verdict counts the 23 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 19 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.