Model comparison
DeepSeek V4 Flash vs GLM-5.2
DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 51.1 on the Noometry Index.
Last verified . 41 shared benchmarks.
DeepSeekSummary
- They share 41 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 2 categories and GLM-5.2 in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 42.3.
- The biggest single-benchmark swing is ARC-AGI-2: 61.4% for DeepSeek V4 Flash and 22.8% for GLM-5.2.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $1.40 / $4.40 for GLM-5.2.
Side by side
| DeepSeek V4 Flash | GLM-5.2 | |
|---|---|---|
| Provider | DeepSeek | Z.ai (Zhipu) |
| Noometry Index | 53.6 | 51.1 |
| Released | 2026-04-24 | 2026-06-13 |
| Weights | Open | Open |
| Context window | 1M | 1M |
| Max output | 393K | 131K |
| Input $ / M tokens | $0.15 | $1.40 |
| Output $ / M tokens | $0.60 | $4.40 |
| Results tracked | 41 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.2 leads
DeepSeek V4 Flash: 47.9 (#59), GLM-5.2: 51.3 (#41)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| FrontierCode | 18.8% | 24.5% |
| LMArena WebDev | 1582 | 1603 |
| SciCode | 49.9% | 50.5% |
| WeirdML | 63% | 70.1% |
| LMArena Coding | 1457 | 1485 |
| ALE-Bench | 1,306 | 1,047 |
| SWE-bench Verified | — | 78.7% |
| DeepSWE | — | 43.8% |
Agentic & Tool Use Not comparable
DeepSeek V4 Flash: —, GLM-5.2: 32.4 (#63)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| APEX-Agents | — | 45.2% |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 31.7% |
| GBAEval | — | 0% |
| Vending-Bench 2 | — | 8,314 |
Reasoning DeepSeek V4 Flash leads
DeepSeek V4 Flash: 53.7 (#30), GLM-5.2: 42.3 (#52)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| ARC-AGI-2 | 61.4% | 22.8% |
| SimpleBench | 61.1% | 58.8% |
| Kagi LLM Benchmark | 52.2% | 62.6% |
| NYT Connections (extended) | 89.6% | 74.3% |
| ARC-AGI-1 | 89% | 77% |
| CritPt | 16.6% | 20.9% |
| Chess Puzzles | 33% | 21% |
| LMArena Hard Prompts | 1444 | 1480 |
| Mystery Game Puzzles | 34% | 19% |
| DTBench | 90.9% | 93.6% |
| LMCA | 41.7% | 45.8% |
| Epoch Capabilities Index | 154.49 | 151.78 |
| EBR-Bench | — | 9.5% |
| Surface Evolver Bench | — | 55.6% |
Math DeepSeek V4 Flash leads
DeepSeek V4 Flash: 60.3 (#37), GLM-5.2: 55.7 (#43)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 57.5% | 59.2% |
| FrontierMath Tier 4 | 24.4% | 29.3% |
| MathArena Final-Answer Competitions | 76.5% | 67.6% |
| OTIS Mock AIME 2024-2025 | 94.4% | 86.4% |
| ProofBench | 56% | 35% |
| LMArena Math | 1427 | 1482 |
Knowledge GLM-5.2 leads
DeepSeek V4 Flash: 55.4 (#48), GLM-5.2: 57.1 (#40)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| GPQA Diamond | 91% | 91.9% |
| SimpleQA Verified | 33.6% | 34.2% |
| LMArena Expert | 1441 | 1486 |
Multilingual GLM-5.2 leads
DeepSeek V4 Flash: 53.0 (#72), GLM-5.2: 55.8 (#26)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| LMArena Non-English | 1420 | 1459 |
| LMArena Chinese | 1468 | 1519 |
| LMArena French | 1439 | 1479 |
| LMArena German | 1418 | 1468 |
| LMArena Japanese | 1406 | 1451 |
| LMArena Korean | 1384 | 1445 |
| LMArena Russian | 1428 | 1466 |
| LMArena Spanish | 1436 | 1477 |
Instruction Following GLM-5.2 leads
DeepSeek V4 Flash: 74.9 (#81), GLM-5.2: 76.9 (#34)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| LMArena Instruction Following | 1421 | 1465 |
Long Context GLM-5.2 leads
DeepSeek V4 Flash: 43.8 (#85), GLM-5.2: 45.3 (#43)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| LMArena Longer Query | 1434 | 1479 |
Writing & Preference GLM-5.2 leads
DeepSeek V4 Flash: 63.8 (#61), GLM-5.2: 70.4 (#21)
| Benchmark | DeepSeek V4 Flash | GLM-5.2 |
|---|---|---|
| LMArena Text | 1432 | 1470 |
| LMArena Creative Writing | 1403 | 1462 |
| EQ-Bench Creative Writing | 1559 | 1757 |
| LMArena Multi-Turn | 1449 | 1469 |
| EQ-Bench 4 | — | 1222 |
Frequently asked questions
Is DeepSeek V4 Flash better than GLM-5.2?
DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 51.1 on the Noometry Index.
Which is cheaper, DeepSeek V4 Flash or GLM-5.2?
DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; GLM-5.2 lists at $1.40 and $4.40.
Is DeepSeek V4 Flash or GLM-5.2 better for coding?
GLM-5.2 scores higher on coding benchmarks: 51.3 versus 47.9 in the Noometry coding category.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do DeepSeek V4 Flash and GLM-5.2 share?
41 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and GLM-5.2 has 51.