Model comparison
Kimi K2.6 vs Mercury 2.5
Kimi K2.6 is the stronger model overall, scoring 47.7 to 33.5 on the Noometry Index. Mercury 2.5 costs 25× less per token, which makes it the better buy when Kimi K2.6's lead doesn't matter for your workload.
Last verified . 4 shared benchmarks.
Summary
- They share 4 benchmarks with published results for both. Kimi K2.6 scores higher in 3 categories and Mercury 2.5 in 0 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2.6 leads 57.0 to 23.3.
- The biggest single-benchmark swing is SciCode: 53.5% for Kimi K2.6 and 38.5% for Mercury 2.5.
- Mercury 2.5 is cheaper at $0.04 / $0.15 per million input/output tokens, against $0.95 / $4 for Kimi K2.6.
- Kimi K2.6 accepts more context: 262K tokens versus 260K.
- Kimi K2.6 has downloadable open weights; the other is API-only.
Side by side
| Kimi K2.6 | Mercury 2.5 | |
|---|---|---|
| Provider | Moonshot AI | Inception |
| Noometry Index | 47.7 | 33.5 |
| Released | 2026-04-20 | 2026-09-08 |
| Weights | Open | Proprietary |
| Context window | 262K | 260K |
| Max output | 262K | 66K |
| Input $ / M tokens | $0.95 | $0.04 |
| Output $ / M tokens | $4 | $0.15 |
| Results tracked | 51 | 4 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K2.6 leads
Kimi K2.6: 50.7 (#43), Mercury 2.5: 39.5 (#156)
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| SciCode | 53.5% | 38.5% |
| ALE-Bench | 1,093 | 301.65 |
| SWE-bench Verified | 76.7% | — |
| LMArena WebDev | 1509 | — |
| WeirdML | 55.9% | — |
| LMArena Coding | 1488 | — |
Agentic & Tool Use Not comparable
Kimi K2.6: 21.9 (#137), Mercury 2.5: —
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| OSWorld 2.0 | 4.6% | — |
| ExploitBench | 18.4% | — |
| GBAEval | 0.9% | — |
| GDP.pdf | 12% | — |
| Vending-Bench 2 | 6,205 | — |
Reasoning Kimi K2.6 leads
Kimi K2.6: 40.5 (#55), Mercury 2.5: 22.4 (#193)
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| CritPt | 8% | 0% |
| NYT Connections (extended) | 87.2% | — |
| Chess Puzzles | 26% | — |
| EBR-Bench | 2.4% | — |
| LMArena Hard Prompts | 1470 | — |
| Mystery Game Puzzles | 18% | — |
| DTBench | 90.9% | — |
| LMCA | 37.3% | — |
| Epoch Capabilities Index | 151.05 | — |
Math Kimi K2.6 leads
Kimi K2.6: 57.0 (#41), Mercury 2.5: 23.3 (#272)
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| ProofBench | 16% | 3% |
| FrontierMath (Tiers 1-3) | 57.2% | — |
| FrontierMath Tier 4 | 25.6% | — |
| MathArena Final-Answer Competitions | 72.9% | — |
| OTIS Mock AIME 2024-2025 | 96.1% | — |
| LMArena Math | 1475 | — |
| FrontierMath (Feb 2025 set) | 39% | — |
| FrontierMath Tier 4 (v1) | 14.6% | — |
Knowledge Not comparable
Kimi K2.6: 54.0 (#54), Mercury 2.5: —
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| GPQA Diamond | 90.8% | — |
| SimpleQA Verified | 34.9% | — |
| Vectara Hallucination Rate | 10.8% | — |
| LMArena Expert | 1491 | — |
Multimodal Not comparable
Kimi K2.6: 31.6 (#103), Mercury 2.5: —
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| LMArena Vision | 1283 | — |
| Blueprint-Bench 2 | 3.9% | — |
| Furniture Assembly | 21.7% | — |
| LMArena Document | 1451 | — |
Multilingual Not comparable
Kimi K2.6: 54.9 (#37), Mercury 2.5: —
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| LMArena Non-English | 1446 | — |
| LMArena Chinese | 1521 | — |
| LMArena French | 1471 | — |
| LMArena German | 1450 | — |
| LMArena Japanese | 1443 | — |
| LMArena Korean | 1427 | — |
| LMArena Russian | 1446 | — |
| LMArena Spanish | 1464 | — |
Instruction Following Not comparable
Kimi K2.6: 76.3 (#43), Mercury 2.5: —
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| LMArena Instruction Following | 1451 | — |
Long Context Not comparable
Kimi K2.6: 44.9 (#52), Mercury 2.5: —
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| LMArena Longer Query | 1468 | — |
Writing & Preference Not comparable
Kimi K2.6: 68.5 (#26), Mercury 2.5: —
| Benchmark | Kimi K2.6 | Mercury 2.5 |
|---|---|---|
| LMArena Text | 1455 | — |
| LMArena Creative Writing | 1434 | — |
| EQ-Bench Creative Writing | 1725 | — |
| EQ-Bench 4 | 1202 | — |
| LMArena Multi-Turn | 1453 | — |
Frequently asked questions
Is Kimi K2.6 better than Mercury 2.5?
Kimi K2.6 is the stronger model overall, scoring 47.7 to 33.5 on the Noometry Index. Mercury 2.5 costs 25× less per token, which makes it the better buy when Kimi K2.6's lead doesn't matter for your workload.
Which is cheaper, Kimi K2.6 or Mercury 2.5?
Mercury 2.5 is cheaper. It lists at $0.04 per million input tokens and $0.15 per million output tokens; Kimi K2.6 lists at $0.95 and $4.
Is Kimi K2.6 or Mercury 2.5 better for coding?
Kimi K2.6 scores higher on coding benchmarks: 50.7 versus 39.5 in the Noometry coding category.
Which has the bigger context window?
Kimi K2.6 does, with 262K tokens against 260K.
How many benchmarks do Kimi K2.6 and Mercury 2.5 share?
4 benchmarks have published results for both models. Kimi K2.6 has 51 scored results on Noometry and Mercury 2.5 has 4.