Qwen3.7 Plus Preview vs Qwen3.8 27B
Wins 3 of 7 areas
Reasoning · Facts · Images and charts
Wins 4 of 7 areas
Coding · Agents · Long documents · Following instructions
Qwen3.8 27B wins more areas, narrowly.Qwen3.7 Plus Preview is cheaper and better at images and charts.
Scores updated · 37 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.
- AgentsCarrying out multi-step tasks on its own04Qwen3.8 27B4 of 4 tests
- CodingWriting and fixing software12Qwen3.8 27B2 of 5 tests · 2 ties
- Long documentsFinding answers in very long texts01Qwen3.8 27B1 of 1 test
- Following instructionsDoing exactly what it is asked01Qwen3.8 27B1 of 1 test
- Images and chartsUnderstanding pictures, charts and video31Qwen3.7 Plus Preview3 of 4 tests
- ReasoningHard problems that need careful thinking20Qwen3.7 Plus Preview2 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up20Qwen3.7 Plus Preview2 of 2 tests
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.
Qwen3.7 Plus Preview costs 43% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.8 27B pulls ahead
- Real work tasks from 44 professionsGDPVal+32.7points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+30.5points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+18.8points ahead
Where Qwen3.7 Plus Preview pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+6.9points ahead
- Harder college exam questions with imagesMMMU-Pro+4.2points ahead
- Research-level physics problemsCritPt+3.7points ahead
Every test, side by side
All 37 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.8 27B
- Terminal-Bench 2.1Qwen3.8 27B by 18.86179.8+18.8
- SWE-bench ProQwen3.8 27B by 4.157.661.7+4.1
- SWE-bench MultilingualQwen3.7 Plus Preview by 275.873.8+2
- LiveCodeBench v6tie89.690.3tie
- SciCodetie46.146.6tie
AgentsQwen3.8 27B
- GDPValQwen3.8 27B by 32.713.546.2+32.7
- τ-Bench V3 · BankingQwen3.8 27B by 30.517.548+30.5
- OSWorld-VerifiedQwen3.8 27B by 1173.384.3+11
- Terminal-Bench 4.0Qwen3.8 27B by 4.615.6+4.6
ReasoningQwen3.7 Plus Preview
- CritPtQwen3.7 Plus Preview by 3.79.15.4+3.7
- Humanity's Last ExamQwen3.7 Plus Preview by 1.735.633.9+1.7
- GPQA Diamondtie9090.5tie
FactsQwen3.7 Plus Preview
- AA-Omniscience · AccuracyQwen3.7 Plus Preview by 6.922.515.6+6.9
- AA-Omniscience · Non-hallucinationQwen3.7 Plus Preview by 2.672.369.7+2.6
Images and chartsQwen3.7 Plus Preview
- CharXiv (RQ)Qwen3.8 27B by 4.385.990.2+4.3
- MMMU-ProQwen3.7 Plus Preview by 4.280.576.3+4.2
- OCRBenchv2Qwen3.7 Plus Preview by 3.667.163.5+3.6
- LMArena · VisionQwen3.7 Plus Preview by 10 rating points12801270+10 rating
Following instructionsQwen3.8 27B
- IFBenchQwen3.8 27B by 1.57879.5+1.5
Other results17 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.
- Agents' Last ExamQwen3.8 27B by 29.713.242.9+29.7
- AA Agentic IndexQwen3.8 27B by 26.819.746.5+26.8
- OSWorld 2.0 (partial)Qwen3.8 27B by 26.521.548+26.5
- DeepSWE 1.1Qwen3.8 27B by 25.716.542.2+25.7
- Toolathlon VerifiedQwen3.8 27B by 16.550.667.1+16.5
- BabyVisionQwen3.8 27B by 15.270.485.6+15.2
- Artificial Analysis Coding IndexQwen3.8 27B by 12.255.968.1+12.2
- AA-OmniscienceQwen3.7 Plus Preview by 11.11.1-10+11.1
- AA IntelligenceQwen3.8 27B by 8.525.233.7+8.5
- JobBenchQwen3.8 27B by 5.827.633.4+5.8
- ERQAQwen3.7 Plus Preview by 4.369.865.5+4.3
- MATH-VisionQwen3.8 27B by 4.390.394.6+4.3
- LVBenchQwen3.7 Plus Preview by 3.876.272.4+3.8
- NL2RepoQwen3.8 27B by 1.241.142.3+1.2
- RealWorldQAQwen3.7 Plus Preview by 186.985.9+1
- AndroidWorldtie8181.9tie
- OmniDocBench 1.5tie91.491.1tie
Questions people ask
Which is better, Qwen3.7 Plus Preview or Qwen3.8 27B?
Qwen3.8 27B wins four of the seven areas where both have results: coding, agents, long documents and following instructions. Qwen3.7 Plus Preview wins reasoning, facts and images and charts, and costs 43% less.
Which is better for coding?
Qwen3.8 27B. It wins 2 of the 5 coding tests both models report; Qwen3.7 Plus Preview wins 1, and 2 are ties.
Which is cheaper?
Qwen3.7 Plus Preview costs $0.40 per million input tokens and $1.60 per million output tokens; Qwen3.8 27B costs $0.50 and $3.00. That makes Qwen3.7 Plus Preview about 43% cheaper for the same work.
How do you compare the two?
We use the 37 benchmark tests both models have published scores on. The verdict counts the 20 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 17 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.