Model comparison
DeepSeek-V3.1-Terminus vs Grok-3 mini
DeepSeek-V3.1-Terminus is the stronger model overall, scoring 42.6 to 41.0 on the Noometry Index.
Last verified . 11 shared benchmarks.
DeepSeekSummary
- They share 11 benchmarks with published results for both. DeepSeek-V3.1-Terminus scores higher in 5 categories and Grok-3 mini in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek-V3.1-Terminus leads 25.7 to 13.8.
- DeepSeek-V3.1-Terminus has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3.1-Terminus | Grok-3 mini | |
|---|---|---|
| Provider | DeepSeek | xAI |
| Noometry Index | 42.6 | 41.0 |
| Released | 2025-09-22 | 2025-04-09 |
| Weights | Open | Proprietary |
| Context window | 164K | — |
| Max output | 66K | — |
| Input $ / M tokens | $0.30 | — |
| Output $ / M tokens | $1 | — |
| Results tracked | 16 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 41.5 (#114), Grok-3 mini: 40.4 (#126)
| Benchmark | DeepSeek-V3.1-Terminus | Grok-3 mini |
|---|---|---|
| LMArena Coding | 1426 | 1380 |
| Aider Polyglot | — | 49.3% |
| SciCode | 40.6% | — |
| WeirdML | — | 42.6% |
| ALE-Bench | 745.17 | — |
Reasoning DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 25.7 (#126), Grok-3 mini: 13.8 (#323)
| Benchmark | DeepSeek-V3.1-Terminus | Grok-3 mini |
|---|---|---|
| Kagi LLM Benchmark | 57.4% | 61.3% |
| LMArena Hard Prompts | 1427 | 1374 |
| ARC-AGI-2 | — | 0.4% |
| ARC-AGI-1 | — | 16.5% |
| CritPt | 1.7% | — |
| DTBench | 81.3% | — |
| LMCA | 28.6% | — |
| Epoch Capabilities Index | — | 140.35 |
Math Grok-3 mini leads
DeepSeek-V3.1-Terminus: 35.7 (#138), Grok-3 mini: 40.3 (#78)
| Benchmark | DeepSeek-V3.1-Terminus | Grok-3 mini |
|---|---|---|
| LMArena Math | 1402 | 1386 |
| OTIS Mock AIME 2024-2025 | — | 77.8% |
| Omni-MATH | — | 31.8% |
| MATH Level 5 | — | 90.9% |
| FrontierMath (Feb 2025 set) | — | 5.9% |
Knowledge Not comparable
DeepSeek-V3.1-Terminus: —, Grok-3 mini: 46.3 (#75)
| Benchmark | DeepSeek-V3.1-Terminus | Grok-3 mini |
|---|---|---|
| GPQA Diamond | — | 76.3% |
| MMLU-Pro | — | 79.9% |
| Confabulations | — | 10.8% |
| GPQA (HELM) | — | 67.5% |
| LMArena Expert | — | 1395 |
Multilingual DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 52.3 (#87), Grok-3 mini: 48.3 (#141)
| Benchmark | DeepSeek-V3.1-Terminus | Grok-3 mini |
|---|---|---|
| LMArena Non-English | 1407 | 1352 |
| LMArena Russian | 1434 | 1352 |
| LMArena Chinese | — | 1387 |
| LMArena French | — | 1357 |
| LMArena German | — | 1350 |
| LMArena Japanese | — | 1343 |
| LMArena Korean | — | 1336 |
| LMArena Spanish | — | 1379 |
Instruction Following Grok-3 mini leads
DeepSeek-V3.1-Terminus: 74.4 (#96), Grok-3 mini: 78.7 (#8)
| Benchmark | DeepSeek-V3.1-Terminus | Grok-3 mini |
|---|---|---|
| LMArena Instruction Following | 1406 | 1356 |
| IFEval | — | 95.1% |
Long Context DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 43.5 (#88), Grok-3 mini: 41.0 (#142)
| Benchmark | DeepSeek-V3.1-Terminus | Grok-3 mini |
|---|---|---|
| LMArena Longer Query | 1426 | 1372 |
| Fiction.LiveBench | — | 66.7% |
Writing & Preference DeepSeek-V3.1-Terminus leads
DeepSeek-V3.1-Terminus: 61.3 (#87), Grok-3 mini: 52.7 (#163)
| Benchmark | DeepSeek-V3.1-Terminus | Grok-3 mini |
|---|---|---|
| LMArena Text | 1420 | 1370 |
| LMArena Creative Writing | 1404 | 1342 |
| LMArena Multi-Turn | 1411 | 1355 |
| Short-Story Creative Writing | — | 73.5% |
| WildBench | — | 65.1% |
Frequently asked questions
Is DeepSeek-V3.1-Terminus better than Grok-3 mini?
DeepSeek-V3.1-Terminus is the stronger model overall, scoring 42.6 to 41.0 on the Noometry Index.
Is DeepSeek-V3.1-Terminus or Grok-3 mini better for coding?
DeepSeek-V3.1-Terminus scores higher on coding benchmarks: 41.5 versus 40.4 in the Noometry coding category.
How many benchmarks do DeepSeek-V3.1-Terminus and Grok-3 mini share?
11 benchmarks have published results for both models. DeepSeek-V3.1-Terminus has 16 scored results on Noometry and Grok-3 mini has 35.