GPT-5.4 vs Qwen3.7 Plus Preview
Wins 5 of 8 areas
Coding · Agents · Reasoning · Math · Long documents
Wins 2 of 8 areas
Images and charts · Following instructions
GPT-5.4 is the stronger all-rounder.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 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.
- AgentsCarrying out multi-step tasks on its own41GPT-5.44 of 5 tests
- ReasoningHard problems that need careful thinking30GPT-5.43 of 3 tests
- MathCompetition and research-level math30GPT-5.43 of 3 tests
- CodingWriting and fixing software31GPT-5.43 of 5 tests · 1 tie
- Long documentsFinding answers in very long texts10GPT-5.41 of 1 test
- Images and chartsUnderstanding pictures, charts and video12Qwen3.7 Plus Preview2 of 3 tests
- Following instructionsDoing exactly what it is asked01Qwen3.7 Plus Preview1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
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 89% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where GPT-5.4 pulls ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+44.2points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+28.4points ahead
- Real work tasks from 44 professionsGDPVal+23.9points ahead
Where Qwen3.7 Plus Preview pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+64points ahead
- Multi-step tasks using many real software toolsMCP Atlas+6points ahead
- Fixes real GitHub issues in many programming languagesSWE-bench Multilingual+4.1points 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.
CodingGPT-5.4
- Terminal-Bench 2.1GPT-5.4 by 17.378.361+17.3
- Terminal-Bench HardGPT-5.4 by 10.657.647+10.6
- SciCodeGPT-5.4 by 10.556.646.1+10.5
- SWE-bench MultilingualQwen3.7 Plus Preview by 4.171.775.8+4.1
- SWE-bench Protie57.757.6tie
AgentsGPT-5.4
- GDPValGPT-5.4 by 23.937.313.4+23.9
- τ-Bench V3 · BankingGPT-5.4 by 22.139.617.5+22.1
- AA ApexAgentsGPT-5.4 by 10.933.322.4+10.9
- MCP AtlasQwen3.7 Plus Preview by 667.273.2+6
- OSWorld-VerifiedGPT-5.4 by 1.77573.3+1.7
ReasoningGPT-5.4
- CritPtGPT-5.4 by 14.323.49.1+14.3
- Humanity's Last ExamGPT-5.4 by 8.143.735.6+8.1
- GPQA DiamondGPT-5.4 by 29290+2
FactsEven
- AA-Omniscience · Non-hallucinationQwen3.7 Plus Preview by 648.372.3+64
- AA-Omniscience · AccuracyGPT-5.4 by 28.450.922.5+28.4
Images and chartsQwen3.7 Plus Preview
- CharXiv (RQ)Qwen3.7 Plus Preview by 3.182.885.9+3.1
- LMArena · VisionGPT-5.4 by 24 rating points13011277+24 rating
- MMMU-ProQwen3.7 Plus Preview by 2.178.480.5+2.1
MathGPT-5.4
- FrontierMath Tiers 1-3 (v2)GPT-5.4 by 44.278.634.4+44.2
- IMOAnswerBenchGPT-5.4 by 5.491.486+5.4
- HMMT Feb. 2026GPT-5.4 by 4.897.792.9+4.8
Following instructionsQwen3.7 Plus Preview
- IFBenchQwen3.7 Plus Preview by 4.173.978+4.1
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.
- DeepSWE 1.1GPT-5.4 by 35.55216.5+35.5
- Apex (Pass@1)GPT-5.4 by 31.454.122.7+31.4
- AA Agentic IndexGPT-5.4 by 24.544.219.7+24.5
- BabyVisionQwen3.7 Plus Preview by 20.749.770.4+20.7
- Artificial Analysis Coding IndexGPT-5.4 by 15.271.155.9+15.2
- AA IntelligenceGPT-5.4 by 13.83925.2+13.8
- τ²-Bench Telecom (AA run)Qwen3.7 Plus Preview by 5.987.193+5.9
- Terminal-Bench 2.0GPT-5.4 by 4.875.170.3+4.8
- AA-OmniscienceGPT-5.4 by 4.75.81.1+4.7
- MCPMarkGPT-5.4 by 3.862.558.7+3.8
- OmniDocBench 1.5Qwen3.7 Plus Preview by 2.389.191.4+2.3
- MATH-VisionGPT-5.4 by 1.79290.3+1.7
- MMLU-ProQwen3.7 Plus Preview by 187.588.5+1
- NL2Repotie41.341.1tie
Questions people ask
Which is better, GPT-5.4 or Qwen3.7 Plus Preview?
GPT-5.4 wins five of the eight areas we test: coding, agents, reasoning, math and long documents. Qwen3.7 Plus Preview wins images and charts and following instructions, and costs 89% less. They are level on facts.
Which is better for coding?
GPT-5.4. It wins 3 of the 5 coding tests both models report; Qwen3.7 Plus Preview wins 1, and 1 is a tie.
Which is cheaper?
GPT-5.4 costs $2.50 per million input tokens and $15.00 per million output tokens; Qwen3.7 Plus Preview costs $0.40 and $1.60. That makes Qwen3.7 Plus Preview about 89% 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 23 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.