Model comparison
Inkling vs Kimi K2.5
Kimi K2.5 is the stronger model overall, scoring 48.1 to 44.1 on the Noometry Index.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. Inkling scores higher in 4 categories and Kimi K2.5 in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2.5 leads 51.8 to 31.3.
- The biggest single-benchmark swing is ARC-AGI-2: 36.5% for Inkling and 11.8% for Kimi K2.5.
- Kimi K2.5 is cheaper at $0.45 / $2.25 per million input/output tokens, against $1.87 / $4.68 for Inkling.
- Kimi K2.5 accepts more context: 262K tokens versus 66K.
Side by side
| Inkling | Kimi K2.5 | |
|---|---|---|
| Provider | Thinking Machines Lab | Moonshot AI |
| Noometry Index | 44.1 | 48.1 |
| Released | 2026-07-15 | 2026-01-27 |
| Weights | Open | Open |
| Context window | 66K | 262K |
| Max output | 66K | 262K |
| Input $ / M tokens | $1.87 | $0.45 |
| Output $ / M tokens | $4.68 | $2.25 |
| Results tracked | 41 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.5 leads
Inkling: 34.5 (#234), Kimi K2.5: 48.8 (#53)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| LMArena WebDev | 1413 | 1437 |
| SciCode | 47% | 49% |
| WeirdML | 32.3% | 45.6% |
| LMArena Coding | 1464 | 1474 |
| ALE-Bench | 946 | 821.65 |
| SWE-bench Verified | — | 73.8% |
| FrontierCode | 14% | — |
| SWE-bench Verified (bash only) | — | 70.8% |
| SWE-bench Multilingual | — | 67.3% |
| FrontierSWE | 4.1% | — |
Agentic & Tool Use Kimi K2.5 leads
Inkling: 29.6 (#85), Kimi K2.5: 34.2 (#48)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| Terminal-Bench | — | 43.2% |
| APEX-Agents | 33.8% | — |
| τ²-bench Banking | 25% | — |
| OSWorld | — | 63.3% |
| Vending-Bench 2 | — | 1,198 |
Reasoning Inkling leads
Inkling: 40.4 (#56), Kimi K2.5: 31.2 (#80)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| ARC-AGI-2 | 36.5% | 11.8% |
| SimpleBench | 50% | 46.8% |
| ARC-AGI-1 | 79.5% | 65.3% |
| CritPt | 5.4% | 3.1% |
| Chess Puzzles | 21% | 12% |
| LMArena Hard Prompts | 1451 | 1453 |
| Epoch Capabilities Index | 148.54 | 148.03 |
| Kagi LLM Benchmark | — | 78.5% |
| NYT Connections (extended) | — | 69.9% |
| EnigmaEval | — | 3.4% |
| Thematic Generalization | — | 69.4% |
| DTBench | 87.5% | — |
| LMCA | 37.6% | — |
Math Kimi K2.5 leads
Inkling: 31.3 (#225), Kimi K2.5: 51.8 (#53)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.9% | 92.2% |
| LMArena Math | 1479 | 1470 |
| FrontierMath (Tiers 1-3) | 33.3% | — |
| FrontierMath Tier 4 | 4.9% | — |
| MathArena Final-Answer Competitions | — | 62.3% |
| ProofBench | 0% | — |
| FrontierMath (Feb 2025 set) | — | 27.9% |
| FrontierMath Tier 4 (v1) | — | 4.2% |
Knowledge Inkling leads
Inkling: 55.1 (#49), Kimi K2.5: 53.6 (#56)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| GPQA Diamond | 88.3% | 87.6% |
| SimpleQA Verified | 40.3% | 34.3% |
| LMArena Expert | 1465 | 1466 |
| Humanity's Last Exam | — | 24.4% |
| Vectara Hallucination Rate | — | 14.2% |
Multimodal Not comparable
Inkling: —, Kimi K2.5: 41.1 (#39)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| LMArena Vision | — | 1269 |
| LMArena Document | — | 1430 |
Multilingual Too close to call
Inkling: 54.0 (#52), Kimi K2.5: 53.9 (#53)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| LMArena Non-English | 1434 | 1433 |
| LMArena Chinese | 1490 | 1495 |
| LMArena French | 1458 | 1454 |
| LMArena German | 1446 | 1441 |
| LMArena Japanese | 1429 | 1421 |
| LMArena Korean | 1404 | 1410 |
| LMArena Russian | 1429 | 1435 |
| LMArena Spanish | 1448 | 1450 |
Instruction Following Too close to call
Inkling: 75.1 (#71), Kimi K2.5: 75.3 (#64)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| LMArena Instruction Following | 1426 | 1431 |
Long Context Kimi K2.5 leads
Inkling: 43.8 (#86), Kimi K2.5: 52.1 (#7)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| LMArena Longer Query | 1434 | 1445 |
| Fiction.LiveBench | — | 86.1% |
| CL-bench | — | 19.3% |
| CL-bench Life | — | 13.2% |
Writing & Preference Too close to call
Inkling: 65.2 (#51), Kimi K2.5: 65.1 (#53)
| Benchmark | Inkling | Kimi K2.5 |
|---|---|---|
| LMArena Text | 1441 | 1445 |
| LMArena Creative Writing | 1387 | 1423 |
| EQ-Bench Creative Writing | 1611 | 1579 |
| LMArena Multi-Turn | 1436 | 1444 |
| EQ-Bench 4 | 1226 | — |
Frequently asked questions
Is Inkling better than Kimi K2.5?
Kimi K2.5 is the stronger model overall, scoring 48.1 to 44.1 on the Noometry Index.
Which is cheaper, Inkling or Kimi K2.5?
Kimi K2.5 is cheaper. It lists at $0.45 per million input tokens and $2.25 per million output tokens; Inkling lists at $1.87 and $4.68.
Is Inkling or Kimi K2.5 better for coding?
Kimi K2.5 scores higher on coding benchmarks: 48.8 versus 34.5 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.5 does, with 262K tokens against 66K.
How many benchmarks do Inkling and Kimi K2.5 share?
31 benchmarks have published results for both models. Inkling has 41 scored results on Noometry and Kimi K2.5 has 51.