Model comparison
GLM-5.3-Flash vs Kimi K2.5
GLM-5.3-Flash is the stronger model overall, scoring 51.8 to 48.1 on the Noometry Index.
Last verified . 28 shared benchmarks.
Summary
- They share 28 benchmarks with published results for both. GLM-5.3-Flash scores higher in 9 categories and Kimi K2.5 in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GLM-5.3-Flash leads 48.0 to 31.2.
- The biggest single-benchmark swing is ARC-AGI-2: 65.8% for GLM-5.3-Flash and 11.8% for Kimi K2.5.
- GLM-5.3-Flash is cheaper at $0.15 / $0.50 per million input/output tokens, against $0.45 / $2.25 for Kimi K2.5.
- GLM-5.3-Flash accepts more context: 1M tokens versus 262K.
Side by side
| GLM-5.3-Flash | Kimi K2.5 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | Moonshot AI |
| Noometry Index | 51.8 | 48.1 |
| Released | 2026-08-20 | 2026-01-27 |
| Weights | Open | Open |
| Context window | 1M | 262K |
| Max output | 131K | 262K |
| Input $ / M tokens | $0.15 | $0.45 |
| Output $ / M tokens | $0.50 | $2.25 |
| Results tracked | 40 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3-Flash leads
GLM-5.3-Flash: 53.1 (#31), Kimi K2.5: 48.8 (#53)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| LMArena WebDev | 1609 | 1437 |
| SciCode | 51.6% | 49% |
| LMArena Coding | 1508 | 1474 |
| ALE-Bench | 303.55 | 821.65 |
| SWE-bench Verified | — | 73.8% |
| DeepSWE | 63.4% | — |
| FrontierCode | 31.8% | — |
| SWE-bench Verified (bash only) | — | 70.8% |
| CursorBench | 36.8% | — |
| SWE-bench Multilingual | — | 67.3% |
| FrontierSWE | 18.1% | — |
| WeirdML | — | 45.6% |
Agentic & Tool Use Too close to call
GLM-5.3-Flash: 34.2 (#47), Kimi K2.5: 34.2 (#48)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | — | 43.2% |
| APEX-Agents | 52.8% | — |
| OSWorld | — | 63.3% |
| GDP.pdf | 14% | — |
| Vending-Bench 2 | — | 1,198 |
Reasoning GLM-5.3-Flash leads
GLM-5.3-Flash: 48.0 (#42), Kimi K2.5: 31.2 (#80)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 65.8% | 11.8% |
| ARC-AGI-1 | 91% | 65.3% |
| CritPt | 15.4% | 3.1% |
| Chess Puzzles | 14% | 12% |
| LMArena Hard Prompts | 1491 | 1453 |
| Epoch Capabilities Index | 151.88 | 148.03 |
| SimpleBench | — | 46.8% |
| Kagi LLM Benchmark | — | 78.5% |
| NYT Connections (extended) | — | 69.9% |
| EnigmaEval | — | 3.4% |
| Thematic Generalization | — | 69.4% |
| Mystery Game Puzzles | 8% | — |
| Surface Evolver Bench | 52.5% | — |
| Bench to the Future 3 | 0.15 | — |
Math GLM-5.3-Flash leads
GLM-5.3-Flash: 53.3 (#47), Kimi K2.5: 51.8 (#53)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 93.9% | 92.2% |
| LMArena Math | 1500 | 1470 |
| FrontierMath (Tiers 1-3) | 55.8% | — |
| FrontierMath Tier 4 | 17.1% | — |
| MathArena Final-Answer Competitions | — | 62.3% |
| ProofBench | 21% | — |
| FrontierMath (Feb 2025 set) | — | 27.9% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
Knowledge GLM-5.3-Flash leads
GLM-5.3-Flash: 58.4 (#36), Kimi K2.5: 53.6 (#56)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 90.2% | 87.6% |
| LMArena Expert | 1513 | 1466 |
| Humanity's Last Exam | — | 24.4% |
| SimpleQA Verified | — | 34.3% |
| Vectara Hallucination Rate | — | 14.2% |
Multimodal GLM-5.3-Flash leads
GLM-5.3-Flash: 42.8 (#27), Kimi K2.5: 41.1 (#39)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| LMArena Vision | 1296 | 1269 |
| LMArena Document | — | 1430 |
Multilingual GLM-5.3-Flash leads
GLM-5.3-Flash: 56.0 (#25), Kimi K2.5: 53.9 (#53)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1462 | 1433 |
| LMArena Chinese | 1527 | 1495 |
| LMArena French | 1496 | 1454 |
| LMArena German | 1470 | 1441 |
| LMArena Japanese | 1429 | 1421 |
| LMArena Korean | 1446 | 1410 |
| LMArena Russian | 1469 | 1435 |
| LMArena Spanish | 1471 | 1450 |
Instruction Following GLM-5.3-Flash leads
GLM-5.3-Flash: 77.5 (#20), Kimi K2.5: 75.3 (#64)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1478 | 1431 |
Long Context Kimi K2.5 leads
GLM-5.3-Flash: 45.4 (#39), Kimi K2.5: 52.1 (#7)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| LMArena Longer Query | 1482 | 1445 |
| Fiction.LiveBench | — | 86.1% |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Too close to call
GLM-5.3-Flash: 65.3 (#50), Kimi K2.5: 65.1 (#53)
| Benchmark | GLM-5.3-Flash | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1471 | 1445 |
| LMArena Creative Writing | 1442 | 1423 |
| LMArena Multi-Turn | 1467 | 1444 |
| EQ-Bench Creative Writing | — | 1579 |
Frequently asked questions
Is GLM-5.3-Flash better than Kimi K2.5?
GLM-5.3-Flash is the stronger model overall, scoring 51.8 to 48.1 on the Noometry Index.
Which is cheaper, GLM-5.3-Flash or Kimi K2.5?
GLM-5.3-Flash is cheaper. It lists at $0.15 per million input tokens and $0.50 per million output tokens; Kimi K2.5 lists at $0.45 and $2.25.
Is GLM-5.3-Flash or Kimi K2.5 better for coding?
GLM-5.3-Flash scores higher on coding benchmarks: 53.1 versus 48.8 in the Noometry coding category.
Which has the bigger context window?
GLM-5.3-Flash does, with 1M tokens against 262K.
How many benchmarks do GLM-5.3-Flash and Kimi K2.5 share?
28 benchmarks have published results for both models. GLM-5.3-Flash has 40 scored results on Noometry and Kimi K2.5 has 51.