Model comparison
GPT-5.6 Terra vs MiniMax-M3
GPT-5.6 Terra is the stronger model overall, scoring 59.2 to 43.8 on the Noometry Index. MiniMax-M3 costs 8.6× less per token, which makes it the better buy when GPT-5.6 Terra's lead doesn't matter for your workload.
Last verified . 38 shared benchmarks.
MiniMaxSummary
- They share 38 benchmarks with published results for both. GPT-5.6 Terra scores higher in 10 categories and MiniMax-M3 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.6 Terra leads 81.6 to 40.0.
- The biggest single-benchmark swing is ProofBench: 74% for GPT-5.6 Terra and 18% for MiniMax-M3.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $2 / $12 for GPT-5.6 Terra.
- GPT-5.6 Terra accepts more context: 1.05M tokens versus 1M.
- MiniMax-M3 has downloadable open weights; the other is API-only.
Side by side
| GPT-5.6 Terra | MiniMax-M3 | |
|---|---|---|
| Provider | OpenAI | MiniMax |
| Noometry Index | 59.2 | 43.8 |
| Released | 2026-07-09 | 2026-06-01 |
| Weights | Proprietary | Open |
| Context window | 1.05M | 1M |
| Max output | 128K | 512K |
| Input $ / M tokens | $2 | $0.30 |
| Output $ / M tokens | $12 | $1.20 |
| Results tracked | 52 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Terra leads
GPT-5.6 Terra: 57.7 (#19), MiniMax-M3: 41.8 (#118)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| FrontierCode | 41.3% | 14.7% |
| LMArena WebDev | 1522 | 1482 |
| SciCode | 55% | 47.1% |
| LMArena Coding | 1484 | 1469 |
| ALE-Bench | 1,951 | 640.02 |
| DeepSWE | 69.6% | — |
| CursorBench | 41.3% | — |
| WeirdML | 78.3% | — |
Agentic & Tool Use GPT-5.6 Terra leads
GPT-5.6 Terra: 40.1 (#25), MiniMax-M3: 22.6 (#130)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| APEX-Agents | 58.2% | 37.7% |
| Vending-Bench 2 | 7,343 | 2,158 |
| OSWorld 2.0 | — | 4.6% |
| BALROG | 53.2% | — |
| GBAEval | — | 0.9% |
| GDP.pdf | 24.7% | — |
Reasoning GPT-5.6 Terra leads
GPT-5.6 Terra: 60.7 (#21), MiniMax-M3: 30.1 (#87)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| SimpleBench | 48.9% | 45.8% |
| NYT Connections (extended) | 78.4% | 65.1% |
| CritPt | 30% | 3.7% |
| Chess Puzzles | 54% | 14% |
| LMArena Hard Prompts | 1468 | 1447 |
| Mystery Game Puzzles | 35% | 8% |
| DTBench | 93.3% | 78.9% |
| LMCA | 55% | 33.7% |
| Surface Evolver Bench | 83.8% | 55% |
| Epoch Capabilities Index | 159.62 | 146.95 |
| ARC-AGI-2 | 83.9% | — |
| Kagi LLM Benchmark | 51.3% | — |
| ARC-AGI-1 | 96.5% | — |
| ForecastBench | — | 61.4 |
Math GPT-5.6 Terra leads
GPT-5.6 Terra: 81.6 (#12), MiniMax-M3: 40.0 (#95)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 99.7% | 71.1% |
| ProofBench | 74% | 18% |
| LMArena Math | 1466 | 1429 |
| FrontierMath (Tiers 1-3) | 86% | — |
| FrontierMath Tier 4 | 70.7% | — |
Knowledge GPT-5.6 Terra leads
GPT-5.6 Terra: 61.2 (#30), MiniMax-M3: 58.4 (#35)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 93.3% | 90.9% |
| LMArena Expert | 1492 | 1461 |
| SimpleQA Verified | 43.2% | — |
Multimodal GPT-5.6 Terra leads
GPT-5.6 Terra: 47.3 (#11), MiniMax-M3: 40.2 (#51)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| LMArena Vision | 1271 | 1253 |
| LMArena Document | 1472 | 1435 |
| Blueprint-Bench 2 | 30.8% | — |
| Furniture Assembly | 54.2% | — |
Multilingual GPT-5.6 Terra leads
GPT-5.6 Terra: 54.4 (#44), MiniMax-M3: 53.0 (#75)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1439 | 1420 |
| LMArena Chinese | 1513 | 1463 |
| LMArena French | 1471 | 1447 |
| LMArena German | 1460 | 1426 |
| LMArena Japanese | 1457 | 1381 |
| LMArena Korean | 1425 | 1372 |
| LMArena Russian | 1450 | 1428 |
| LMArena Spanish | 1448 | 1432 |
Instruction Following Too close to call
GPT-5.6 Terra: 76.4 (#40), MiniMax-M3: 75.5 (#62)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1454 | 1433 |
Long Context Too close to call
GPT-5.6 Terra: 44.4 (#68), MiniMax-M3: 44.2 (#72)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1451 | 1445 |
Writing & Preference GPT-5.6 Terra leads
GPT-5.6 Terra: 70.2 (#23), MiniMax-M3: 62.1 (#83)
| Benchmark | GPT-5.6 Terra | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1447 | 1433 |
| LMArena Creative Writing | 1410 | 1404 |
| EQ-Bench 4 | 1234 | 1150 |
| LMArena Multi-Turn | 1449 | 1442 |
| EQ-Bench Creative Writing | 1855 | — |
Frequently asked questions
Is GPT-5.6 Terra better than MiniMax-M3?
GPT-5.6 Terra is the stronger model overall, scoring 59.2 to 43.8 on the Noometry Index. MiniMax-M3 costs 8.6× less per token, which makes it the better buy when GPT-5.6 Terra's lead doesn't matter for your workload.
Which is cheaper, GPT-5.6 Terra or MiniMax-M3?
MiniMax-M3 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; GPT-5.6 Terra lists at $2 and $12.
Is GPT-5.6 Terra or MiniMax-M3 better for coding?
GPT-5.6 Terra scores higher on coding benchmarks: 57.7 versus 41.8 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Terra does, with 1.05M tokens against 1M.
How many benchmarks do GPT-5.6 Terra and MiniMax-M3 share?
38 benchmarks have published results for both models. GPT-5.6 Terra has 52 scored results on Noometry and MiniMax-M3 has 41.