Model comparison
Mercury vs Qwen2.5 72B Instruct
Mercury is the stronger model overall, scoring 37.6 to 31.9 on the Noometry Index.
Last verified . 8 shared benchmarks.
Alibaba (Qwen)Summary
- They share 8 benchmarks with published results for both. Mercury scores higher in 2 categories and Qwen2.5 72B Instruct in 4 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in coding, where Mercury leads 38.7 to 33.2.
- Qwen2.5 72B Instruct has downloadable open weights; the other is API-only.
Side by side
| Mercury | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | Inception | Alibaba (Qwen) |
| Noometry Index | 37.6 | 31.9 |
| Released | — | 2024-09 |
| Weights | Proprietary | Open |
| Context window | — | 131K |
| Max output | — | 8K |
| Input $ / M tokens | — | $1.40 |
| Output $ / M tokens | — | $5.60 |
| Results tracked | 9 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Mercury leads
Mercury: 38.7 (#170), Qwen2.5 72B Instruct: 33.2 (#260)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Coding | 1322 | 1292 |
| WeirdML | — | 16% |
| BigCodeBench Instruct | — | 45.8% |
| BigCodeBench Complete | — | 55.9% |
Agentic & Tool Use Not comparable
Mercury: —, Qwen2.5 72B Instruct: 22.1 (#133)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning Qwen2.5 72B Instruct leads
Mercury: 17.5 (#293), Qwen2.5 72B Instruct: 22.3 (#199)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1285 | 1271 |
| Kagi LLM Benchmark | 21.6% | — |
| DTBench | — | 62.9% |
| LMCA | — | 13.4% |
| BIG-Bench Hard | — | 79.8% |
| Epoch Capabilities Index | — | 129 |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math Not comparable
Mercury: —, Qwen2.5 72B Instruct: 19.3 (#287)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 8.1% |
| Omni-MATH | — | 33% |
| LMArena Math | — | 1283 |
| MATH Level 5 | — | 63.2% |
Knowledge Not comparable
Mercury: —, Qwen2.5 72B Instruct: 27.0 (#253)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | — | 49.1% |
| MMLU-Pro | — | 63.1% |
| Confabulations | — | 19.1% |
| GPQA (HELM) | — | 42.6% |
| LMArena Expert | — | 1245 |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multilingual Too close to call
Mercury: 41.6 (#206), Qwen2.5 72B Instruct: 41.0 (#213)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1260 | 1252 |
| LMArena Chinese | — | 1272 |
| LMArena French | — | 1280 |
| LMArena German | — | 1234 |
| LMArena Japanese | — | 1180 |
| LMArena Korean | — | 1188 |
| LMArena Russian | — | 1264 |
| LMArena Spanish | — | 1256 |
Instruction Following Too close to call
Mercury: 65.2 (#224), Qwen2.5 72B Instruct: 65.5 (#221)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1239 | 1254 |
| IFEval | — | 80.6% |
Long Context Too close to call
Mercury: 38.4 (#198), Qwen2.5 72B Instruct: 38.9 (#188)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1266 | 1282 |
Writing & Preference Too close to call
Mercury: 46.2 (#221), Qwen2.5 72B Instruct: 46.7 (#215)
| Benchmark | Mercury | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1282 | 1269 |
| LMArena Creative Writing | 1191 | 1221 |
| LMArena Multi-Turn | 1282 | 1272 |
| WildBench | — | 80.2% |
Frequently asked questions
Is Mercury better than Qwen2.5 72B Instruct?
Mercury is the stronger model overall, scoring 37.6 to 31.9 on the Noometry Index.
Is Mercury or Qwen2.5 72B Instruct better for coding?
Mercury scores higher on coding benchmarks: 38.7 versus 33.2 in the Noometry coding category.
How many benchmarks do Mercury and Qwen2.5 72B Instruct share?
8 benchmarks have published results for both models. Mercury has 9 scored results on Noometry and Qwen2.5 72B Instruct has 43.