Qwen3.5 397B A17B vs Qwen3.7 Plus Preview
Wins 1 of 8 areas
Long documents
Wins 5 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math
Qwen3.7 Plus Preview is the stronger all-rounder.Qwen3.5 397B A17B is better at long documents.
Scores updated · 43 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 software07Qwen3.7 Plus Preview7 of 7 tests
- MathCompetition and research-level math03Qwen3.7 Plus Preview3 of 3 tests
- AgentsCarrying out multi-step tasks on its own13Qwen3.7 Plus Preview3 of 4 tests
- ReasoningHard problems that need careful thinking02Qwen3.7 Plus Preview2 of 3 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video02Qwen3.7 Plus Preview2 of 2 tests
- Long documentsFinding answers in very long texts10Qwen3.5 397B A17B1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
- Following instructionsDoing exactly what it is asked00Even0 each · 1 tie
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 52% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Qwen3.7 Plus Preview pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+61.2points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+9.7points ahead
- Research-level physics problemsCritPt+7.4points ahead
Where Qwen3.5 397B A17B pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+8.3points ahead
- Reasons across sets of long documentsAA-LCR+4.3points ahead
- Real work tasks from 44 professionsGDPVal+1.3points ahead
Every test, side by side
All 43 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingQwen3.7 Plus Preview
- Terminal-Bench 2.1Qwen3.7 Plus Preview by 9.751.361+9.7
- SWE-bench ProQwen3.7 Plus Preview by 6.750.957.6+6.7
- SWE-bench MultilingualQwen3.7 Plus Preview by 6.569.375.8+6.5
- Terminal-Bench HardQwen3.7 Plus Preview by 6.140.947+6.1
- LiveCodeBench v6Qwen3.7 Plus Preview by 683.689.6+6
- SciCodeQwen3.7 Plus Preview by 1.344.846.1+1.3
- SWE-bench VerifiedQwen3.7 Plus Preview by 1.376.477.7+1.3
AgentsQwen3.7 Plus Preview
- AA ApexAgentsQwen3.7 Plus Preview by 7.115.322.4+7.1
- τ-Bench V3 · BankingQwen3.7 Plus Preview by 4.113.417.5+4.1
- GDPValQwen3.5 397B A17B by 1.314.813.5+1.3
- Terminal-Bench 4.0Qwen3.7 Plus Preview by 101+1
ReasoningQwen3.7 Plus Preview
- CritPtQwen3.7 Plus Preview by 7.41.79.1+7.4
- Humanity's Last ExamQwen3.7 Plus Preview by 6.62935.6+6.6
- GPQA Diamondtie89.390tie
FactsEven
- AA-Omniscience · Non-hallucinationQwen3.7 Plus Preview by 61.211.172.3+61.2
- AA-Omniscience · AccuracyQwen3.5 397B A17B by 8.330.822.5+8.3
Images and chartsQwen3.7 Plus Preview
- MMMU-ProQwen3.7 Plus Preview by 3.277.380.5+3.2
- LMArena · VisionQwen3.7 Plus Preview by 16 rating points12641280+16 rating
MathQwen3.7 Plus Preview
- IMOAnswerBenchQwen3.7 Plus Preview by 5.180.986+5.1
- HMMT Feb. 2026Qwen3.7 Plus Preview by 587.992.9+5
- FrontierMath Tiers 1-3 (v2)Qwen3.7 Plus Preview by 3.231.234.4+3.2
Other results20 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.
- AA-OmniscienceQwen3.7 Plus Preview by 31.8-30.81.1+31.8
- DeepPlanningQwen3.7 Plus Preview by 2834.362.3+28
- Terminal-Bench 2.0Qwen3.7 Plus Preview by 17.852.570.3+17.8
- MCPMarkQwen3.7 Plus Preview by 12.646.158.7+12.6
- QwenClawBenchQwen3.7 Plus Preview by 1051.861.8+10
- Toolathlon VerifiedQwen3.7 Plus Preview by 9.940.750.6+9.9
- AA Agentic IndexQwen3.7 Plus Preview by 9.110.619.7+9.1
- NL2RepoQwen3.7 Plus Preview by 8.932.241.1+8.9
- Artificial Analysis Coding IndexQwen3.7 Plus Preview by 7.748.255.9+7.7
- AA IntelligenceQwen3.7 Plus Preview by 6.818.425.2+6.8
- VITA-BenchQwen3.5 397B A17B by 4.149.745.6+4.1
- IncludeQwen3.5 397B A17B by 2.685.683+2.6
- τ²-Bench Telecom (AA run)Qwen3.5 397B A17B by 2.695.693+2.6
- IFEvalQwen3.7 Plus Preview by 292.694.6+2
- SuperGPQAQwen3.7 Plus Preview by 170.471.4+1
- MMLU-Protie87.888.5tie
- MMLU-ProXtie84.785.4tie
- MMMLUtie88.589tie
- MMLU-Reduxtie94.994.5tie
- BFCLv4tie72.972.9tie
Questions people ask
Which is better, Qwen3.5 397B A17B or Qwen3.7 Plus Preview?
Qwen3.7 Plus Preview wins five of the eight areas we test: coding, agents, reasoning, images and charts and math. Qwen3.5 397B A17B wins long documents. They are level on facts and following instructions.
Which is better for coding?
Qwen3.7 Plus Preview. It wins 7 of the 7 coding tests both models report; Qwen3.5 397B A17B wins none.
Which is cheaper?
Qwen3.5 397B A17B costs $0.60 per million input tokens and $3.60 per million output tokens; Qwen3.7 Plus Preview costs $0.40 and $1.60. That makes Qwen3.7 Plus Preview about 52% cheaper for the same work.
How do you compare the two?
We use the 43 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 20 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.