DeepSeek-V3.2 vs GLM-4.7-Flash
Wins 4 of 6 areas
Coding · Reasoning · Facts · Long documents
Wins 0 of 6 areas
—
DeepSeek-V3.2 is the stronger all-rounder.GLM-4.7-Flash is cheaper.
Scores updated · 25 tests both models report · How we compare
Where each one wins
Tests won in each of the six 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 software40DeepSeek-V3.24 of 4 tests
- ReasoningHard problems that need careful thinking30DeepSeek-V3.23 of 3 tests
- FactsGetting facts right instead of making them up30DeepSeek-V3.23 of 3 tests
- Long documentsFinding answers in very long texts10DeepSeek-V3.21 of 1 test
- AgentsCarrying out multi-step tasks on its own11Even1 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.
GLM-4.7-Flash costs 33% 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 DeepSeek-V3.2 pulls ahead
- Reasons across sets of long documentsAA-LCR+31.6points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+25.9points ahead
- Recent programming contest problemsLiveCodeBench v6+19.3points ahead
Where GLM-4.7-Flash pulls ahead
- Real work tasks from 44 professionsGDPVal+7.1points ahead
Every test, side by side
All 25 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-V3.2
- LiveCodeBench v6DeepSeek-V3.2 by 19.383.364+19.3
- SWE-bench VerifiedDeepSeek-V3.2 by 13.973.159.2+13.9
- Terminal-Bench HardDeepSeek-V3.2 by 13.635.622+13.6
- SciCodeDeepSeek-V3.2 by 5.238.933.7+5.2
AgentsEven
- BrowseCompDeepSeek-V3.2 by 8.651.442.8+8.6
- GDPValGLM-4.7-Flash by 7.19.816.9+7.1
ReasoningDeepSeek-V3.2
- GPQA DiamondDeepSeek-V3.2 by 25.98458.1+25.9
- Humanity's Last ExamDeepSeek-V3.2 by 1724.67.6+17
- CritPtDeepSeek-V3.2 by 2.62.90.3+2.6
FactsDeepSeek-V3.2
- AA-Omniscience · AccuracyDeepSeek-V3.2 by 16.83316.1+16.8
- AA-Omniscience · Non-hallucinationDeepSeek-V3.2 by 11.217.36.1+11.2
- Vectara HHEM hallucination ratelower is betterDeepSeek-V3.2 by 36.39.3+3
Other results11 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-OmniscienceDeepSeek-V3.2 by 40.1-22.5-62.6+40.1
- AA Agentic IndexGLM-4.7-Flash by 27.718.346+27.7
- Artificial Analysis Coding IndexDeepSeek-V3.2 by 18.344.225.9+18.3
- vectara_avg_summary_lengthGLM-4.7-Flash by 9.86271.8+9.8
- τ²-Bench Telecom (AA run)GLM-4.7-Flash by 8.290.698.8+8.2
- AA IntelligenceDeepSeek-V3.2 by 6.621.514.9+6.6
- GPQA (unspecified)DeepSeek-V3.2 by 4.779.975.2+4.7
- vectara_factual_consistencyDeepSeek-V3.2 by 393.790.7+3
- AIME 2025DeepSeek-V3.2 by 1.593.191.6+1.5
- vectara_answer_rateDeepSeek-V3.2 by 192.691.6+1
- τ²-Benchtie80.379.5tie
Questions people ask
Which is better, DeepSeek-V3.2 or GLM-4.7-Flash?
DeepSeek-V3.2 wins four of the six areas where both have results: coding, reasoning, facts and long documents. GLM-4.7-Flash wins none, but costs 33% less. They are level on agents and following instructions.
Which is better for coding?
DeepSeek-V3.2. It wins 4 of the 4 coding tests both models report; GLM-4.7-Flash wins none.
Which is cheaper?
DeepSeek-V3.2 costs $0.28 per million input tokens and $0.42 per million output tokens; GLM-4.7-Flash costs $0.07 and $0.40. That makes GLM-4.7-Flash about 33% cheaper for the same work.
How do you compare the two?
We use the 25 benchmark tests both models have published scores on. The verdict counts the 14 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 11 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.