Up to four models
GLM-5.3-Flash vs Muse Spark 1.1
On the index (GPT-4 = 100), GLM-5.3-Flash scores 161 and Muse Spark 1.1 161. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
161145–177
WeightsOpen
ReleasedAug 2026
Price in / out$0.15 / $0.50
Context1.05M
Core tests taken6 of 9
161145–177
WeightsClosed
ReleasedJul 2026
Price in / out$1.25 / $4.25
Context1.05M
Core tests taken2 of 9
GLM-5.3-Flash and Muse Spark 1.1 are too close to call (0 points apart, within the uncertainty), ahead on 2 of 2 shared core tests, and costs 8.4× less.
Test by testthird-party results
Every result as a table
| Test | GLM-5.3-Flash | Muse Spark 1.1 |
|---|---|---|
| Reasoning & knowledge | ||
| LiveBench Data Analysis | 76.4% | 72.5% |
| LiveBench IF | 52.8% | 69.6% |
| LiveBench Language | 77.3% | 74.3% |
| LiveBench Reasoning | 77.6% | 87.7% |
| GPQA Diamond (Vals AI) | 86.4% | 91.2% |
| MMLU Pro | 86.1% | 88.7% |
| MMMU Pro | 86.0% | 86.6% |
| Maths | ||
| LiveBench Mathematics | 81.2% | 87.1% |
| Chess puzzles | 21.0% | 34.0% |
| Surface Evolver Bench | 52.5% | 52.5% |
| Coding | ||
| LiveBench Agentic Coding | 56.8% | 58.5% |
| LiveBench Coding | 79.0% | 77.2% |
| Code Migration | 20.5% | 31.1% |
| LiveCodeBench | 80.5% | 85.9% |
| ProgramBench | 0.0% | 0.0% |
| SWE-bench | 92.0% | 82.0% |
| Vibe Code Bench v1.1 | 30.8% | 72.2% |
| DeepSWE | 63.4% | 53.3% |
| Agents | ||
| APEX-Agents | 52.8% | 31.8% |
| SkillsBench | 40.2% | 59.2% |
| Professional work | ||
| Finance Agent (v2) | 57.9% | 57.2% |
| EMB | 55.9% | 56.4% |
| LegalBench | 83.9% | 85.0% |
| Legal Research Bench | 45.2% | 38.0% |
| MedScribe | 88.9% | 88.9% |
| TaxEval v2 | 75.6% | 79.7% |
Tests at least two of these models have taken. Bold rows are the core tests.