Model comparison
Hy3 vs MiniMax-M2.5
Hy3 is the stronger model overall, scoring 44.2 to 38.3 on the Noometry Index.
Last verified . 19 shared benchmarks.
Tencent
MiniMaxSummary
- They share 19 benchmarks with published results for both. Hy3 scores higher in 7 categories and MiniMax-M2.5 in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Hy3 leads 40.1 to 26.9.
- The biggest single-benchmark swing is NYT Connections (extended): 41.2% for Hy3 and 16.8% for MiniMax-M2.5.
- Hy3 is cheaper at $0.0825 / $0.33 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.5.
- Hy3 accepts more context: 262K tokens versus 205K.
Side by side
| Hy3 | MiniMax-M2.5 | |
|---|---|---|
| Provider | Tencent | MiniMax |
| Noometry Index | 44.2 | 38.3 |
| Released | 2026-07-06 | 2026-02-12 |
| Weights | Open | Open |
| Context window | 262K | 205K |
| Max output | 128K | 131K |
| Input $ / M tokens | $0.0825 | $0.30 |
| Output $ / M tokens | $0.33 | $1.20 |
| Results tracked | 19 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2.5 leads
Hy3: 46.8 (#63), MiniMax-M2.5: 48.1 (#58)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| LMArena WebDev | 1508 | 1387 |
| LMArena Coding | 1464 | 1381 |
| SWE-bench Verified (bash only) | — | 75.8% |
| SWE-bench Multilingual | — | 68.3% |
| ALE-Bench | — | 618.17 |
Agentic & Tool Use Not comparable
Hy3: —, MiniMax-M2.5: 30.4 (#77)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| Terminal-Bench | — | 42.7% |
| Vending-Bench 2 | — | -23.16 |
Reasoning Hy3 leads
Hy3: 26.1 (#136), MiniMax-M2.5: 17.5 (#292)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| NYT Connections (extended) | 41.2% | 16.8% |
| LMArena Hard Prompts | 1447 | 1372 |
| ARC-AGI-2 | — | 4.9% |
| Kagi LLM Benchmark | — | 55.2% |
| ARC-AGI-1 | — | 63.7% |
| Epoch Capabilities Index | — | 146.68 |
Math Hy3 leads
Hy3: 40.1 (#93), MiniMax-M2.5: 26.9 (#253)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Math | 1475 | 1378 |
| ProofBench | — | 4% |
Knowledge Hy3 leads
Hy3: 40.8 (#114), MiniMax-M2.5: 39.2 (#135)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Expert | 1460 | 1379 |
| Vectara Hallucination Rate | — | 9.1% |
Multilingual Hy3 leads
Hy3: 53.5 (#65), MiniMax-M2.5: 47.1 (#152)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Non-English | 1426 | 1338 |
| LMArena Chinese | 1493 | 1393 |
| LMArena French | 1461 | 1362 |
| LMArena German | 1439 | 1362 |
| LMArena Japanese | 1392 | 1171 |
| LMArena Korean | 1395 | 1232 |
| LMArena Russian | 1432 | 1358 |
| LMArena Spanish | 1456 | 1354 |
Instruction Following Hy3 leads
Hy3: 75.1 (#70), MiniMax-M2.5: 71.5 (#148)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Instruction Following | 1426 | 1353 |
Long Context Hy3 leads
Hy3: 44.1 (#75), MiniMax-M2.5: 37.5 (#216)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Longer Query | 1442 | 1366 |
| CL-bench | — | 11.4% |
| CL-bench Life | — | 6.3% |
Writing & Preference Hy3 leads
Hy3: 62.2 (#81), MiniMax-M2.5: 53.9 (#153)
| Benchmark | Hy3 | MiniMax-M2.5 |
|---|---|---|
| LMArena Text | 1439 | 1359 |
| LMArena Creative Writing | 1402 | 1331 |
| LMArena Multi-Turn | 1436 | 1364 |
| EQ-Bench Creative Writing | — | 1361 |
Frequently asked questions
Is Hy3 better than MiniMax-M2.5?
Hy3 is the stronger model overall, scoring 44.2 to 38.3 on the Noometry Index.
Which is cheaper, Hy3 or MiniMax-M2.5?
Hy3 is cheaper. It lists at $0.0825 per million input tokens and $0.33 per million output tokens; MiniMax-M2.5 lists at $0.30 and $1.20.
Is Hy3 or MiniMax-M2.5 better for coding?
MiniMax-M2.5 scores higher on coding benchmarks: 48.1 versus 46.8 in the Noometry coding category.
Which has the bigger context window?
Hy3 does, with 262K tokens against 205K.
How many benchmarks do Hy3 and MiniMax-M2.5 share?
19 benchmarks have published results for both models. Hy3 has 19 scored results on Noometry and MiniMax-M2.5 has 33.