Model comparison
DeepSeek-V3 vs Trinity Large Thinking
DeepSeek-V3 and Trinity Large Thinking score almost the same on the Noometry Index (39.5 vs 38.6), so choose on price, context window or the category you care about most.
Last verified . 20 shared benchmarks.
DeepSeek
Arcee AISummary
- They share 20 benchmarks with published results for both. DeepSeek-V3 scores higher in 5 categories and Trinity Large Thinking in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where DeepSeek-V3 leads 42.3 to 34.1.
- Both cost about the same: $0.24 input and $0.90 output per million tokens.
- Trinity Large Thinking accepts more context: 262K tokens versus 164K.
Side by side
| DeepSeek-V3 | Trinity Large Thinking | |
|---|---|---|
| Provider | DeepSeek | Arcee AI |
| Noometry Index | 39.5 | 38.6 |
| Released | 2024-12-26 | 2026-04-01 |
| Weights | Open | Open |
| Context window | 164K | 262K |
| Max output | 164K | 80K |
| Input $ / M tokens | $0.24 | $0.25 |
| Output $ / M tokens | $0.90 | $0.80 |
| Results tracked | 60 | 24 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3 leads
DeepSeek-V3: 42.3 (#106), Trinity Large Thinking: 34.1 (#244)
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| SciCode | 35.8% | 36.1% |
| LMArena Coding | 1368 | 1381 |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1238 |
| WeirdML | 36.1% | — |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, Trinity Large Thinking: —
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| METR Time Horizons | 49.6% | — |
Reasoning DeepSeek-V3 leads
DeepSeek-V3: 20.5 (#236), Trinity Large Thinking: 16.9 (#298)
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| CritPt | 0% | 0.9% |
| LMArena Hard Prompts | 1365 | 1350 |
| SimpleBench | 27.2% | — |
| Kagi LLM Benchmark | 52.3% | — |
| NYT Connections (extended) | — | 16.5% |
| Thematic Generalization | — | 41.6% |
| LiveBench Reasoning | 65.8% | — |
| DTBench | 64.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| LMCA | 15.5% | — |
| Surface Evolver Bench | — | 15.6% |
| BIG-Bench Hard | 87.5% | — |
| Epoch Capabilities Index | 135.94 | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math Trinity Large Thinking leads
DeepSeek-V3: 32.1 (#219), Trinity Large Thinking: 37.6 (#149)
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| LMArena Math | 1373 | 1366 |
| OTIS Mock AIME 2024-2025 | 37.8% | — |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge Trinity Large Thinking leads
DeepSeek-V3: 37.5 (#155), Trinity Large Thinking: 40.9 (#113)
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| Vectara Hallucination Rate | 6.1% | 6.9% |
| LMArena Expert | 1351 | 1360 |
| GPQA Diamond | 67.6% | — |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multilingual DeepSeek-V3 leads
DeepSeek-V3: 48.5 (#143), Trinity Large Thinking: 46.2 (#160)
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| LMArena Non-English | 1358 | 1325 |
| LMArena Chinese | 1391 | 1373 |
| LMArena French | 1385 | 1374 |
| LMArena German | 1374 | 1356 |
| LMArena Japanese | 1333 | 1311 |
| LMArena Korean | 1319 | 1306 |
| LMArena Russian | 1373 | 1337 |
| LMArena Spanish | 1358 | 1357 |
Instruction Following DeepSeek-V3 leads
DeepSeek-V3: 72.8 (#130), Trinity Large Thinking: 70.5 (#162)
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| LMArena Instruction Following | 1345 | 1334 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context Trinity Large Thinking leads
DeepSeek-V3: 34.0 (#253), Trinity Large Thinking: 41.3 (#144)
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| LMArena Longer Query | 1352 | 1355 |
| Fiction.LiveBench | 50% | — |
Writing & Preference DeepSeek-V3 leads
DeepSeek-V3: 57.4 (#130), Trinity Large Thinking: 53.8 (#158)
| Benchmark | DeepSeek-V3 | Trinity Large Thinking |
|---|---|---|
| LMArena Text | 1375 | 1340 |
| LMArena Creative Writing | 1364 | 1320 |
| LMArena Multi-Turn | 1389 | 1342 |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than Trinity Large Thinking?
DeepSeek-V3 and Trinity Large Thinking score almost the same on the Noometry Index (39.5 vs 38.6), so choose on price, context window or the category you care about most.
Which is cheaper, DeepSeek-V3 or Trinity Large Thinking?
Trinity Large Thinking is cheaper. It lists at $0.25 per million input tokens and $0.80 per million output tokens; DeepSeek-V3 lists at $0.24 and $0.90.
Is DeepSeek-V3 or Trinity Large Thinking better for coding?
DeepSeek-V3 scores higher on coding benchmarks: 42.3 versus 34.1 in the Noometry coding category.
Which has the bigger context window?
Trinity Large Thinking does, with 262K tokens against 164K.
How many benchmarks do DeepSeek-V3 and Trinity Large Thinking share?
20 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and Trinity Large Thinking has 24.