Up to four models
GPT-6 Astra vs Muse Spark Open
On the index (GPT-4 = 100), GPT-6 Astra scores 199 and Muse Spark Open 167. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
GPT-6 Astra
199182–215
WeightsClosed
ReleasedSep 2026
Price in / out$10 / $50
Context1.05M
Core tests taken7 of 9
Muse Spark Open167151–184
WeightsOpen
ReleasedOct 2026
Price in / out$1.25 / $4.25
Context1.05M
Core tests taken6 of 9
GPT-6 Astra leads Muse Spark Open by 31 points, ahead on 5 of 6 shared core tests, at 10× the price.
Test by testthird-party results
GPT-6 AstraMuse Spark Open
Muse Spark OpenEvery result as a table
| Test | GPT-6 Astra | Muse Spark Open |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 95.8% | 88.6% |
| SimpleQA Verified | 75.6% | 56.6% |
| Maths | ||
| FrontierMath (tiers 1-3) | 93.7% | 69.8% |
| AIME-style maths (OTIS mock) | 100.0% | 98.1% |
| FrontierMath (tier 4) | 97.6% | 31.7% |
| Coding | ||
| Terminal-Bench 4.0 | 59.6% | 6.6% |
| Code Migration | 67.7% | 21.2% |
| Vibe Code Bench v1.1 | 89.6% | 76.2% |
| Agents | ||
| APEX-Agents | 64.7% | 33.4% |
| Professional work | ||
| Finance Agent (v2) | 53.5% | 61.3% |
| EMB | 71.7% | 53.7% |
| HLAB-Vals-1.1 | 7.1% | 25.0% |
| Legal Research Bench | 39.4% | 38.5% |
| Tax Agent Bench | 20.7% | 26.3% |
| Novel problems | ||
| ARC-AGI-2 | 95.0% | 57.5% |
| ARC-AGI-1 | 98.5% | 89.0% |
Tests at least two of these models have taken. Bold rows are the core tests.
