Model comparison
DeepSeek V4 Pro vs Trinity Large Thinking
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 38.6 on the Noometry Index. Trinity Large Thinking costs 2.6× less per token, which makes it the better buy when DeepSeek V4 Pro's lead doesn't matter for your workload.
Last verified . 23 shared benchmarks.
DeepSeek
Arcee AISummary
- They share 23 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 8 categories and Trinity Large Thinking in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 16.9.
- The biggest single-benchmark swing is NYT Connections (extended): 91.3% for DeepSeek V4 Pro and 16.5% for Trinity Large Thinking.
- Trinity Large Thinking is cheaper at $0.25 / $0.80 per million input/output tokens, against $0.66 / $1.98 for DeepSeek V4 Pro.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 262K.
Side by side
| DeepSeek V4 Pro | Trinity Large Thinking | |
|---|---|---|
| Provider | DeepSeek | Arcee AI |
| Noometry Index | 54.3 | 38.6 |
| Released | 2026-04-24 | 2026-04-01 |
| Weights | Open | Open |
| Context window | 1M | 262K |
| Max output | 393K | 80K |
| Input $ / M tokens | $0.66 | $0.25 |
| Output $ / M tokens | $1.98 | $0.80 |
| Results tracked | 48 | 24 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.4 (#34), Trinity Large Thinking: 34.1 (#244)
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| LMArena WebDev | 1582 | 1238 |
| SciCode | 51% | 36.1% |
| LMArena Coding | 1470 | 1381 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| WeirdML | 66.2% | — |
| ALE-Bench | 1,403 | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Pro: 32.8 (#58), Trinity Large Thinking: —
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), Trinity Large Thinking: 16.9 (#298)
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| NYT Connections (extended) | 91.3% | 16.5% |
| CritPt | 18% | 0.9% |
| LMArena Hard Prompts | 1461 | 1350 |
| Surface Evolver Bench | 40% | 15.6% |
| ARC-AGI-2 | 61.3% | — |
| Kagi LLM Benchmark | 53.5% | — |
| ARC-AGI-1 | 90.5% | — |
| Chess Puzzles | 47% | — |
| Thematic Generalization | — | 41.6% |
| Mystery Game Puzzles | 43% | — |
| DTBench | 93.9% | — |
| LMCA | 45.5% | — |
| Epoch Capabilities Index | 155.31 | — |
| ForecastBench | 56.1 | — |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Trinity Large Thinking: 37.6 (#149)
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| LMArena Math | 1455 | 1366 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| OTIS Mock AIME 2024-2025 | 98.6% | — |
| ProofBench | 50% | — |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#31), Trinity Large Thinking: 40.9 (#113)
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| Vectara Hallucination Rate | 8.6% | 6.9% |
| LMArena Expert | 1464 | 1360 |
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.4 (#45), Trinity Large Thinking: 46.2 (#160)
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| LMArena Non-English | 1439 | 1325 |
| LMArena Chinese | 1486 | 1373 |
| LMArena French | 1472 | 1374 |
| LMArena German | 1458 | 1356 |
| LMArena Japanese | 1445 | 1311 |
| LMArena Korean | 1447 | 1306 |
| LMArena Russian | 1453 | 1337 |
| LMArena Spanish | 1458 | 1357 |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.1 (#47), Trinity Large Thinking: 70.5 (#162)
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| LMArena Instruction Following | 1448 | 1334 |
Long Context DeepSeek V4 Pro leads
DeepSeek V4 Pro: 45.0 (#51), Trinity Large Thinking: 41.3 (#144)
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| LMArena Longer Query | 1458 | 1355 |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.5 (#46), Trinity Large Thinking: 53.8 (#158)
| Benchmark | DeepSeek V4 Pro | Trinity Large Thinking |
|---|---|---|
| LMArena Text | 1451 | 1340 |
| LMArena Creative Writing | 1446 | 1320 |
| LMArena Multi-Turn | 1467 | 1342 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Trinity Large Thinking?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 38.6 on the Noometry Index. Trinity Large Thinking costs 2.6× less per token, which makes it the better buy when DeepSeek V4 Pro's lead doesn't matter for your workload.
Which is cheaper, DeepSeek V4 Pro 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 V4 Pro lists at $0.66 and $1.98.
Is DeepSeek V4 Pro or Trinity Large Thinking better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 34.1 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 262K.
How many benchmarks do DeepSeek V4 Pro and Trinity Large Thinking share?
23 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Trinity Large Thinking has 24.