Model comparison
DeepSeek-V3 vs GPT-5.6 Terra
GPT-5.6 Terra is the stronger model overall, scoring 59.2 to 39.5 on the Noometry Index. DeepSeek-V3 costs 11× less per token, which makes it the better buy when GPT-5.6 Terra's lead doesn't matter for your workload.
Last verified . 28 shared benchmarks.
DeepSeekSummary
- They share 28 benchmarks with published results for both. DeepSeek-V3 scores higher in 0 categories and GPT-5.6 Terra in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.6 Terra leads 81.6 to 32.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for DeepSeek-V3 and 99.7% for GPT-5.6 Terra.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $2 / $12 for GPT-5.6 Terra.
- GPT-5.6 Terra accepts more context: 1.05M tokens versus 164K.
- DeepSeek-V3 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-V3 | GPT-5.6 Terra | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 39.5 | 59.2 |
| Released | 2024-12-26 | 2026-07-09 |
| Weights | Open | Proprietary |
| Context window | 164K | 1.05M |
| Max output | 164K | 128K |
| Input $ / M tokens | $0.24 | $2 |
| Output $ / M tokens | $0.90 | $12 |
| Results tracked | 60 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Terra leads
DeepSeek-V3: 42.3 (#106), GPT-5.6 Terra: 57.7 (#19)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| SciCode | 35.8% | 55% |
| WeirdML | 36.1% | 78.3% |
| LMArena Coding | 1368 | 1484 |
| DeepSWE | — | 69.6% |
| FrontierCode | — | 41.3% |
| Aider Polyglot | 55.1% | — |
| CursorBench | — | 41.3% |
| LMArena WebDev | — | 1522 |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| ALE-Bench | — | 1,951 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, GPT-5.6 Terra: 40.1 (#25)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| APEX-Agents | — | 58.2% |
| BALROG | — | 53.2% |
| GDP.pdf | — | 24.7% |
| METR Time Horizons | 49.6% | — |
| Vending-Bench 2 | — | 7,343 |
Reasoning GPT-5.6 Terra leads
DeepSeek-V3: 20.5 (#236), GPT-5.6 Terra: 60.7 (#21)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| SimpleBench | 27.2% | 48.9% |
| Kagi LLM Benchmark | 52.3% | 51.3% |
| CritPt | 0% | 30% |
| LMArena Hard Prompts | 1365 | 1468 |
| DTBench | 64.8% | 93.3% |
| LMCA | 15.5% | 55% |
| Epoch Capabilities Index | 135.94 | 159.62 |
| ARC-AGI-2 | — | 83.9% |
| NYT Connections (extended) | — | 78.4% |
| ARC-AGI-1 | — | 96.5% |
| Chess Puzzles | — | 54% |
| LiveBench Reasoning | 65.8% | — |
| Mystery Game Puzzles | — | 35% |
| LiveBench Data Analysis | 60.9% | — |
| Surface Evolver Bench | — | 83.8% |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math GPT-5.6 Terra leads
DeepSeek-V3: 32.1 (#219), GPT-5.6 Terra: 81.6 (#12)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 99.7% |
| LMArena Math | 1373 | 1466 |
| FrontierMath (Tiers 1-3) | — | 86% |
| FrontierMath Tier 4 | — | 70.7% |
| ProofBench | — | 74% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge GPT-5.6 Terra leads
DeepSeek-V3: 37.5 (#155), GPT-5.6 Terra: 61.2 (#30)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| GPQA Diamond | 67.6% | 93.3% |
| LMArena Expert | 1351 | 1492 |
| SimpleQA Verified | — | 43.2% |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| Vectara Hallucination Rate | 6.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multimodal Not comparable
DeepSeek-V3: —, GPT-5.6 Terra: 47.3 (#11)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Vision | — | 1271 |
| Blueprint-Bench 2 | — | 30.8% |
| Furniture Assembly | — | 54.2% |
| LMArena Document | — | 1472 |
Multilingual GPT-5.6 Terra leads
DeepSeek-V3: 48.5 (#143), GPT-5.6 Terra: 54.4 (#44)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Non-English | 1358 | 1439 |
| LMArena Chinese | 1391 | 1513 |
| LMArena French | 1385 | 1471 |
| LMArena German | 1374 | 1460 |
| LMArena Japanese | 1333 | 1457 |
| LMArena Korean | 1319 | 1425 |
| LMArena Russian | 1373 | 1450 |
| LMArena Spanish | 1358 | 1448 |
Instruction Following GPT-5.6 Terra leads
DeepSeek-V3: 72.8 (#130), GPT-5.6 Terra: 76.4 (#40)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Instruction Following | 1345 | 1454 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context GPT-5.6 Terra leads
DeepSeek-V3: 34.0 (#253), GPT-5.6 Terra: 44.4 (#68)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Longer Query | 1352 | 1451 |
| Fiction.LiveBench | 50% | — |
Writing & Preference GPT-5.6 Terra leads
DeepSeek-V3: 57.4 (#130), GPT-5.6 Terra: 70.2 (#23)
| Benchmark | DeepSeek-V3 | GPT-5.6 Terra |
|---|---|---|
| LMArena Text | 1375 | 1447 |
| LMArena Creative Writing | 1364 | 1410 |
| EQ-Bench Creative Writing | 1472 | 1855 |
| LMArena Multi-Turn | 1389 | 1449 |
| Short-Story Creative Writing | 77% | — |
| WildBench | 83% | — |
| EQ-Bench 4 | — | 1234 |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than GPT-5.6 Terra?
GPT-5.6 Terra is the stronger model overall, scoring 59.2 to 39.5 on the Noometry Index. DeepSeek-V3 costs 11× less per token, which makes it the better buy when GPT-5.6 Terra's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-V3 or GPT-5.6 Terra?
DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; GPT-5.6 Terra lists at $2 and $12.
Is DeepSeek-V3 or GPT-5.6 Terra better for coding?
GPT-5.6 Terra scores higher on coding benchmarks: 57.7 versus 42.3 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Terra does, with 1.05M tokens against 164K.
How many benchmarks do DeepSeek-V3 and GPT-5.6 Terra share?
28 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and GPT-5.6 Terra has 52.