Model comparison
GPT-5.4 vs Muse Spark 1.3
GPT-5.4 is the stronger model overall, scoring 59.4 to 54.8 on the Noometry Index. Muse Spark 1.3 costs 2.8× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.
Last verified . 34 shared benchmarks.
MetaSummary
- They share 34 benchmarks with published results for both. GPT-5.4 scores higher in 6 categories and Muse Spark 1.3 in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-5.4 leads 65.3 to 42.6.
- The biggest single-benchmark swing is Mystery Game Puzzles: 37% for GPT-5.4 and 25% for Muse Spark 1.3.
- Muse Spark 1.3 is cheaper at $1.25 / $4.25 per million input/output tokens, against $2.50 / $15 for GPT-5.4.
- GPT-5.4 accepts more context: 1.05M tokens versus 1.05M.
Side by side
| GPT-5.4 | Muse Spark 1.3 | |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 59.4 | 54.8 |
| Released | 2026-03-05 | 2026-09-02 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.05M |
| Max output | 128K | 131K |
| Input $ / M tokens | $2.50 | $1.25 |
| Output $ / M tokens | $15 | $4.25 |
| Results tracked | 68 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Muse Spark 1.3 leads
GPT-5.4: 52.6 (#33), Muse Spark 1.3: 56.6 (#21)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| LMArena WebDev | 1465 | 1657 |
| SciCode | 56.6% | 59.7% |
| LMArena Coding | 1497 | 1514 |
| SWE-bench Verified | 76.9% | — |
| DeepSWE | 51.8% | — |
| CursorBench | — | 41.6% |
| GSO | 31.4% | — |
| WeirdML | 77.7% | — |
| MirrorCode | 15.6% | — |
| ALE-Bench | 1,607 | — |
| AlgoTune | 1.85 | — |
Agentic & Tool Use GPT-5.4 leads
GPT-5.4: 46.5 (#13), Muse Spark 1.3: 38.6 (#30)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| APEX-Agents | 52.4% | 57.8% |
| Terminal-Bench | 81.8% | — |
| τ²-bench Banking | 39.4% | — |
| DeepResearch Bench | 35.1% | — |
| PostTrainBench | 19% | — |
| GBAEval | 45.1% | — |
| GDP.pdf | — | 27.6% |
| LMArena Search | 1197 | — |
| METR Time Horizons | 74.3% | — |
| Vending-Bench 2 | 6,144 | — |
Reasoning GPT-5.4 leads
GPT-5.4: 61.8 (#19), Muse Spark 1.3: 54.0 (#27)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| NYT Connections (extended) | 91.3% | 85.1% |
| CritPt | 23.4% | 26% |
| Chess Puzzles | 44% | 38% |
| LMArena Hard Prompts | 1485 | 1503 |
| Mystery Game Puzzles | 37% | 25% |
| DTBench | 94.4% | 96.5% |
| LMCA | 52% | 53.9% |
| Epoch Capabilities Index | 156.81 | 156.75 |
| ARC-AGI-2 | 74% | — |
| Kagi LLM Benchmark | 63.8% | — |
| ARC-AGI-1 | 93.7% | — |
| EnigmaEval | 16% | — |
| Thematic Generalization | 80% | — |
| EBR-Bench | 25.4% | — |
| Bench to the Future 3 | — | 0.14 |
| ForecastBench | 59.5 | — |
Math Too close to call
GPT-5.4: 73.5 (#19), Muse Spark 1.3: 73.1 (#21)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 78.6% | 74.4% |
| FrontierMath Tier 4 | 49% | 46.3% |
| OTIS Mock AIME 2024-2025 | 97.8% | 99.2% |
| ProofBench | 56% | 58% |
| LMArena Math | 1488 | 1494 |
| MathArena Final-Answer Competitions | 83.1% | — |
| FrontierMath (Feb 2025 set) | 47.6% | — |
| FrontierMath Tier 4 (v1) | 27.1% | — |
Knowledge GPT-5.4 leads
GPT-5.4: 65.3 (#14), Muse Spark 1.3: 42.6 (#95)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| LMArena Expert | 1507 | 1516 |
| GPQA Diamond | 93.3% | — |
| Humanity's Last Exam | 36.2% | — |
| SimpleQA Verified | 45.1% | — |
| Vectara Hallucination Rate | 7% | — |
Multimodal Too close to call
GPT-5.4: 43.7 (#20), Muse Spark 1.3: 43.7 (#22)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| LMArena Vision | 1303 | 1309 |
| LMArena Document | 1471 | 1471 |
| Blueprint-Bench 2 | 27.1% | — |
| Furniture Assembly | 37.5% | — |
Multilingual Muse Spark 1.3 leads
GPT-5.4: 56.2 (#23), Muse Spark 1.3: 57.4 (#8)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| LMArena Non-English | 1465 | 1481 |
| LMArena Chinese | 1519 | 1529 |
| LMArena French | 1493 | 1524 |
| LMArena German | 1472 | 1515 |
| LMArena Japanese | 1485 | 1474 |
| LMArena Korean | 1448 | 1501 |
| LMArena Russian | 1480 | 1490 |
| LMArena Spanish | 1454 | 1490 |
Instruction Following Too close to call
GPT-5.4: 77.1 (#27), Muse Spark 1.3: 77.5 (#22)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| LMArena Instruction Following | 1469 | 1477 |
Long Context GPT-5.4 leads
GPT-5.4: 50.3 (#8), Muse Spark 1.3: 45.6 (#32)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| LMArena Longer Query | 1473 | 1488 |
| CL-bench | 27.9% | — |
| CL-bench Life | 21.7% | — |
Writing & Preference Muse Spark 1.3 leads
GPT-5.4: 71.9 (#17), Muse Spark 1.3: 73.6 (#9)
| Benchmark | GPT-5.4 | Muse Spark 1.3 |
|---|---|---|
| LMArena Text | 1469 | 1490 |
| LMArena Creative Writing | 1439 | 1455 |
| EQ-Bench Creative Writing | 1840 | 1906 |
| LMArena Multi-Turn | 1482 | 1482 |
| EQ-Bench 4 | 1272 | — |
Frequently asked questions
Is GPT-5.4 better than Muse Spark 1.3?
GPT-5.4 is the stronger model overall, scoring 59.4 to 54.8 on the Noometry Index. Muse Spark 1.3 costs 2.8× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.
Which is cheaper, GPT-5.4 or Muse Spark 1.3?
Muse Spark 1.3 is cheaper. It lists at $1.25 per million input tokens and $4.25 per million output tokens; GPT-5.4 lists at $2.50 and $15.
Is GPT-5.4 or Muse Spark 1.3 better for coding?
Muse Spark 1.3 scores higher on coding benchmarks: 56.6 versus 52.6 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 does, with 1.05M tokens against 1.05M.
How many benchmarks do GPT-5.4 and Muse Spark 1.3 share?
34 benchmarks have published results for both models. GPT-5.4 has 68 scored results on Noometry and Muse Spark 1.3 has 37.