Model comparison
Llama-3.3-70B-Instruct vs MiMo-V2.5-Pro
MiMo-V2.5-Pro is the stronger model overall, scoring 45.2 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 3.5× less per token, which makes it the better buy when MiMo-V2.5-Pro's lead doesn't matter for your workload.
Last verified . 21 shared benchmarks.
MetaSummary
- They share 21 benchmarks with published results for both. Llama-3.3-70B-Instruct scores higher in 0 categories and MiMo-V2.5-Pro in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where MiMo-V2.5-Pro leads 40.0 to 15.3.
- The biggest single-benchmark swing is DTBench: 59.5% for Llama-3.3-70B-Instruct and 84.5% for MiMo-V2.5-Pro.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $0.43 / $0.87 for MiMo-V2.5-Pro.
- MiMo-V2.5-Pro accepts more context: 1.05M tokens versus 128K.
Side by side
| Llama-3.3-70B-Instruct | MiMo-V2.5-Pro | |
|---|---|---|
| Provider | Meta | Xiaomi |
| Noometry Index | 30.6 | 45.2 |
| Released | 2024-12-06 | 2026-04-22 |
| Weights | Open | Open |
| Context window | 128K | 1.05M |
| Max output | 4K | 131K |
| Input $ / M tokens | $0.10 | $0.43 |
| Output $ / M tokens | $0.32 | $0.87 |
| Results tracked | 43 | 27 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiMo-V2.5-Pro leads
Llama-3.3-70B-Instruct: 31.0 (#290), MiMo-V2.5-Pro: 47.4 (#60)
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| SciCode | 26% | 50.2% |
| LMArena Coding | 1268 | 1503 |
| LMArena WebDev | — | 1479 |
| WeirdML | 14.4% | — |
| BigCodeBench Instruct | 46.9% | — |
| LiveBench Coding | 36.6% | — |
| BigCodeBench Complete | 57.5% | — |
| ALE-Bench | — | 899.8 |
Agentic & Tool Use Not comparable
Llama-3.3-70B-Instruct: 25.8 (#105), MiMo-V2.5-Pro: —
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| Berkeley Function Calling Leaderboard | 31.9% | — |
| BALROG | 23% | — |
Reasoning MiMo-V2.5-Pro leads
Llama-3.3-70B-Instruct: 14.1 (#327), MiMo-V2.5-Pro: 26.8 (#130)
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| CritPt | 0% | 4% |
| LMArena Hard Prompts | 1257 | 1488 |
| DTBench | 59.5% | 84.5% |
| LMCA | 17.5% | 29.5% |
| SimpleBench | 19.9% | — |
| NYT Connections (extended) | — | 34.4% |
| LiveBench Reasoning | 50.8% | — |
| LiveBench Data Analysis | 49.5% | — |
| Epoch Capabilities Index | 127.33 | — |
| ForecastBench | 58.6 | — |
| LiveBench | 50.2% | — |
Math MiMo-V2.5-Pro leads
Llama-3.3-70B-Instruct: 15.3 (#298), MiMo-V2.5-Pro: 40.0 (#96)
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Math | 1267 | 1481 |
| OTIS Mock AIME 2024-2025 | 5.1% | — |
| ProofBench | — | 22% |
| LiveBench Math | 42.2% | — |
| MATH Level 5 | 41.6% | — |
Knowledge MiMo-V2.5-Pro leads
Llama-3.3-70B-Instruct: 30.6 (#226), MiMo-V2.5-Pro: 42.2 (#98)
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Expert | 1225 | 1503 |
| GPQA Diamond | 47.4% | — |
| Confabulations | 22.8% | — |
| Vectara Hallucination Rate | 4.1% | — |
| MMLU | 86.3% | — |
Multilingual MiMo-V2.5-Pro leads
Llama-3.3-70B-Instruct: 39.9 (#220), MiMo-V2.5-Pro: 55.1 (#34)
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Non-English | 1236 | 1449 |
| LMArena Chinese | 1217 | 1507 |
| LMArena French | 1281 | 1488 |
| LMArena German | 1251 | 1458 |
| LMArena Japanese | 1150 | 1412 |
| LMArena Korean | 1143 | 1437 |
| LMArena Russian | 1252 | 1450 |
| LMArena Spanish | 1270 | 1471 |
Instruction Following MiMo-V2.5-Pro leads
Llama-3.3-70B-Instruct: 71.1 (#157), MiMo-V2.5-Pro: 77.5 (#21)
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Instruction Following | 1242 | 1477 |
| LiveBench Instruction Following | 82.7% | — |
Long Context MiMo-V2.5-Pro leads
Llama-3.3-70B-Instruct: 26.4 (#295), MiMo-V2.5-Pro: 45.4 (#37)
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Longer Query | 1256 | 1483 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference MiMo-V2.5-Pro leads
Llama-3.3-70B-Instruct: 47.6 (#207), MiMo-V2.5-Pro: 65.3 (#49)
| Benchmark | Llama-3.3-70B-Instruct | MiMo-V2.5-Pro |
|---|---|---|
| LMArena Text | 1274 | 1465 |
| LMArena Creative Writing | 1250 | 1440 |
| LMArena Multi-Turn | 1280 | 1477 |
| EQ-Bench Creative Writing | — | 1493 |
| EQ-Bench 4 | — | 1208 |
| LiveBench Language | 39.2% | — |
Frequently asked questions
Is Llama-3.3-70B-Instruct better than MiMo-V2.5-Pro?
MiMo-V2.5-Pro is the stronger model overall, scoring 45.2 to 30.6 on the Noometry Index. Llama-3.3-70B-Instruct costs 3.5× less per token, which makes it the better buy when MiMo-V2.5-Pro's lead doesn't matter for your workload.
Which is cheaper, Llama-3.3-70B-Instruct or MiMo-V2.5-Pro?
Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; MiMo-V2.5-Pro lists at $0.43 and $0.87.
Is Llama-3.3-70B-Instruct or MiMo-V2.5-Pro better for coding?
MiMo-V2.5-Pro scores higher on coding benchmarks: 47.4 versus 31.0 in the Noometry coding category.
Which has the bigger context window?
MiMo-V2.5-Pro does, with 1.05M tokens against 128K.
How many benchmarks do Llama-3.3-70B-Instruct and MiMo-V2.5-Pro share?
21 benchmarks have published results for both models. Llama-3.3-70B-Instruct has 43 scored results on Noometry and MiMo-V2.5-Pro has 27.