Model comparison
Kimi K2.5 vs Muse Spark 1.2
Muse Spark 1.2 is the stronger model overall, scoring 49.8 to 47.8 on the Noometry Index. Kimi K2.5 costs 2.2× less per token, which makes it the better buy when Muse Spark 1.2's lead doesn't matter for your workload.
Last verified . 21 shared benchmarks.
MetaSummary
- They share 21 benchmarks with published results for both. Kimi K2.5 scores higher in 4 categories and Muse Spark 1.2 in 6 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Muse Spark 1.2 leads 50.4 to 31.2.
- The biggest single-benchmark swing is SimpleBench: 46.8% for Kimi K2.5 and 74.5% for Muse Spark 1.2.
- Kimi K2.5 is cheaper at $0.45 / $2.25 per million input/output tokens, against $1.25 / $4.25 for Muse Spark 1.2.
- Muse Spark 1.2 accepts more context: 1.05M tokens versus 262K.
- Kimi K2.5 has downloadable open weights; the other is API-only.
Side by side
| Kimi K2.5 | Muse Spark 1.2 | |
|---|---|---|
| Provider | Moonshot AI | Meta |
| Noometry Index | 47.8 | 49.8 |
| Released | 2026-01-27 | 2026-08-05 |
| Weights | Open | Proprietary |
| Context window | 262K | 1.05M |
| Max output | 262K | 131K |
| Input $ / M tokens | $0.45 | $1.25 |
| Output $ / M tokens | $2.25 | $4.25 |
| Results tracked | 51 | 28 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Kimi K2.5: 48.7 (#48), Muse Spark 1.2: 48.6 (#49)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| LMArena WebDev | 1436 | 1532 |
| SciCode | 49% | 56.4% |
| WeirdML | 45.6% | 60.3% |
| LMArena Coding | 1474 | 1497 |
| SWE-bench Verified | 73.8% | — |
| DeepSWE | — | 54.9% |
| SWE-bench Verified (bash only) | 70.8% | — |
| SWE-bench Multilingual | 67.3% | — |
| FrontierSWE | — | 12% |
| ALE-Bench | 821.65 | — |
Agentic & Tool Use Kimi K2.5 leads
Kimi K2.5: 34.0 (#44), Muse Spark 1.2: 29.5 (#77)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| Terminal-Bench | 43.2% | — |
| APEX-Agents | — | 36.4% |
| OSWorld | 63.3% | — |
| GDP.pdf | — | 16% |
| Vending-Bench 2 | 1,198 | — |
Reasoning Muse Spark 1.2 leads
Kimi K2.5: 31.2 (#74), Muse Spark 1.2: 50.4 (#32)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| SimpleBench | 46.8% | 74.5% |
| NYT Connections (extended) | 69.9% | 79.2% |
| CritPt | 3.1% | 17.7% |
| LMArena Hard Prompts | 1454 | 1490 |
| Epoch Capabilities Index | 148.05 | 155.18 |
| ARC-AGI-2 | 11.8% | — |
| Kagi LLM Benchmark | 78.5% | — |
| ARC-AGI-1 | 65.3% | — |
| Chess Puzzles | 12% | — |
| EnigmaEval | 3.4% | — |
| Thematic Generalization | 69.4% | — |
| DTBench | — | 94.7% |
| LMCA | — | 48.4% |
Math Kimi K2.5 leads
Kimi K2.5: 50.3 (#46), Muse Spark 1.2: 43.8 (#68)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| MathArena Final-Answer Competitions | 62.3% | — |
| OTIS Mock AIME 2024-2025 | 92.2% | — |
| ProofBench | — | 43% |
| LMArena Math | 1470 | — |
| FrontierMath (Feb 2025 set) | 27.9% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |
Knowledge Too close to call
Kimi K2.5: 53.6 (#52), Muse Spark 1.2: 53.9 (#49)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| SimpleQA Verified | 34.3% | 60.3% |
| LMArena Expert | 1466 | 1485 |
| GPQA Diamond | 87.6% | — |
| Humanity's Last Exam | 24.4% | — |
| Vectara Hallucination Rate | 14.2% | — |
Multimodal Muse Spark 1.2 leads
Kimi K2.5: 40.4 (#37), Muse Spark 1.2: 42.6 (#24)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| LMArena Vision | 1269 | 1304 |
| LMArena Document | 1430 | — |
Multilingual Muse Spark 1.2 leads
Kimi K2.5: 54.1 (#50), Muse Spark 1.2: 57.4 (#10)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| LMArena Non-English | 1433 | 1479 |
| LMArena Chinese | 1495 | 1519 |
| LMArena Russian | 1436 | 1494 |
| LMArena French | 1454 | — |
| LMArena German | 1441 | — |
| LMArena Japanese | 1422 | — |
| LMArena Korean | 1412 | — |
| LMArena Spanish | 1450 | — |
Instruction Following Muse Spark 1.2 leads
Kimi K2.5: 75.5 (#60), Muse Spark 1.2: 77.1 (#30)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| LMArena Instruction Following | 1431 | 1466 |
Long Context Kimi K2.5 leads
Kimi K2.5: 52.1 (#7), Muse Spark 1.2: 45.4 (#35)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| LMArena Longer Query | 1445 | 1483 |
| Fiction.LiveBench | 86.1% | — |
| CL-bench | 19.3% | — |
| CL-bench Life | 13.2% | — |
Writing & Preference Muse Spark 1.2 leads
Kimi K2.5: 65.3 (#51), Muse Spark 1.2: 72.6 (#14)
| Benchmark | Kimi K2.5 | Muse Spark 1.2 |
|---|---|---|
| LMArena Text | 1445 | 1486 |
| LMArena Creative Writing | 1423 | 1444 |
| EQ-Bench Creative Writing | 1579 | 1840 |
| LMArena Multi-Turn | 1444 | 1504 |
Frequently asked questions
Is Kimi K2.5 better than Muse Spark 1.2?
Muse Spark 1.2 is the stronger model overall, scoring 49.8 to 47.8 on the Noometry Index. Kimi K2.5 costs 2.2× less per token, which makes it the better buy when Muse Spark 1.2's lead doesn't matter for your workload.
Which is cheaper, Kimi K2.5 or Muse Spark 1.2?
Kimi K2.5 is cheaper. It lists at $0.45 per million input tokens and $2.25 per million output tokens; Muse Spark 1.2 lists at $1.25 and $4.25.
Is Kimi K2.5 or Muse Spark 1.2 better for coding?
They score almost the same on coding (48.7 vs 48.6); test both on your own repository before choosing.
Which has the bigger context window?
Muse Spark 1.2 does, with 1.05M tokens against 262K.
How many benchmarks do Kimi K2.5 and Muse Spark 1.2 share?
21 benchmarks have published results for both models. Kimi K2.5 has 51 scored results on Noometry and Muse Spark 1.2 has 28.