Model comparison
DeepSeek-R1 vs GPT-5.6 Sol
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 42.3 on the Noometry Index. DeepSeek-R1 costs 8.8× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Last verified . 31 shared benchmarks.
DeepSeekSummary
- They share 31 benchmarks with published results for both. DeepSeek-R1 scores higher in 1 category and GPT-5.6 Sol in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.6 Sol leads 74.8 to 18.6.
- The biggest single-benchmark swing is ARC-AGI-2: 1.3% for DeepSeek-R1 and 92.5% for GPT-5.6 Sol.
- DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $4 / $20 for GPT-5.6 Sol.
- GPT-5.6 Sol accepts more context: 1.05M tokens versus 164K.
Side by side
| DeepSeek-R1 | GPT-5.6 Sol | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 42.3 | 65.0 |
| Released | 2025-01-20 | 2026-07-09 |
| Weights | Proprietary | Proprietary |
| Context window | 164K | 1.05M |
| Max output | 64K | 128K |
| Input $ / M tokens | $0.50 | $4 |
| Output $ / M tokens | $2.15 | $20 |
| Results tracked | 52 | 65 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Sol leads
DeepSeek-R1: 46.3 (#68), GPT-5.6 Sol: 65.1 (#7)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| SciCode | 35.7% | 57.1% |
| WeirdML | 41.6% | 89.4% |
| LMArena Coding | 1427 | 1498 |
| ALE-Bench | 804.12 | 2,177 |
| DeepSWE | — | 72.7% |
| FrontierCode | — | 47.5% |
| Aider Polyglot | 71.4% | — |
| CursorBench | — | 41.7% |
| LMArena WebDev | — | 1618 |
| FrontierSWE | — | 32.2% |
| GSO | — | 76.5% |
| LiveBench Coding | 66.7% | — |
| MirrorCode | — | 20% |
| AlgoTune | 1.7 | — |
Agentic & Tool Use GPT-5.6 Sol leads
DeepSeek-R1: 30.7 (#75), GPT-5.6 Sol: 50.3 (#7)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| BALROG | 34.9% | 60% |
| APEX-Agents | — | 51.4% |
| OSWorld 2.0 | — | 27.3% |
| τ²-bench Banking | — | 46.9% |
| DeepResearch Bench | 35.1% | — |
| PostTrainBench | — | 36.2% |
| GBAEval | — | 52.6% |
| GDP.pdf | — | 30.7% |
| LMArena Search | — | 1257 |
| METR Time Horizons | 53.8% | — |
| Vending-Bench 2 | — | 9,619 |
Reasoning GPT-5.6 Sol leads
DeepSeek-R1: 18.6 (#278), GPT-5.6 Sol: 74.8 (#8)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| ARC-AGI-2 | 1.3% | 92.5% |
| SimpleBench | 40.8% | 71.7% |
| Kagi LLM Benchmark | 69.4% | 67% |
| ARC-AGI-1 | 21.2% | 97.5% |
| CritPt | 1.1% | 32.3% |
| LMArena Hard Prompts | 1416 | 1484 |
| Epoch Capabilities Index | 141.29 | 161.66 |
| NYT Connections (extended) | — | 93.8% |
| Chess Puzzles | — | 64% |
| EnigmaEval | — | 37.1% |
| EBR-Bench | — | 44.8% |
| LiveBench Reasoning | 83.2% | — |
| Mystery Game Puzzles | — | 58% |
| DTBench | — | 96% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 59.2% |
| Surface Evolver Bench | — | 93.1% |
| Bench to the Future 3 | — | 0.14 |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math GPT-5.6 Sol leads
DeepSeek-R1: 43.8 (#79), GPT-5.6 Sol: 85.6 (#9)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 100% |
| LMArena Math | 1400 | 1474 |
| FrontierMath (Tiers 1-3) | — | 89.1% |
| FrontierMath Tier 4 | — | 82.9% |
| ProofBench | — | 83% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
| FrontierMath Erdős | — | 0% |
Knowledge GPT-5.6 Sol leads
DeepSeek-R1: 44.5 (#87), GPT-5.6 Sol: 64.3 (#18)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| GPQA Diamond | 76.3% | 93.5% |
| Vectara Hallucination Rate | 11.3% | 12.4% |
| LMArena Expert | 1394 | 1516 |
| SimpleQA Verified | — | 69.7% |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, GPT-5.6 Sol: 48.6 (#9)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| LMArena Vision | — | 1281 |
| Blueprint-Bench 2 | — | 33.6% |
| Furniture Assembly | — | 56.7% |
| LMArena Document | — | 1483 |
Multilingual GPT-5.6 Sol leads
DeepSeek-R1: 52.4 (#85), GPT-5.6 Sol: 55.3 (#32)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| LMArena Non-English | 1412 | 1452 |
| LMArena Chinese | 1442 | 1527 |
| LMArena French | 1417 | 1477 |
| LMArena German | 1404 | 1476 |
| LMArena Japanese | 1391 | 1471 |
| LMArena Korean | 1360 | 1442 |
| LMArena Russian | 1423 | 1468 |
| LMArena Spanish | 1411 | 1441 |
Instruction Following GPT-5.6 Sol leads
DeepSeek-R1: 72.0 (#143), GPT-5.6 Sol: 77.7 (#16)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| LMArena Instruction Following | 1382 | 1482 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context Too close to call
DeepSeek-R1: 45.4 (#36), GPT-5.6 Sol: 45.4 (#42)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| LMArena Longer Query | 1391 | 1480 |
| Fiction.LiveBench | 75% | — |
Writing & Preference GPT-5.6 Sol leads
DeepSeek-R1: 61.4 (#88), GPT-5.6 Sol: 73.3 (#12)
| Benchmark | DeepSeek-R1 | GPT-5.6 Sol |
|---|---|---|
| LMArena Text | 1428 | 1457 |
| LMArena Creative Writing | 1405 | 1448 |
| EQ-Bench Creative Writing | 1500 | 1972 |
| LMArena Multi-Turn | 1405 | 1460 |
| Short-Story Creative Writing | 83% | — |
| WildBench | 82.8% | — |
| EQ-Bench 4 | — | 1250 |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than GPT-5.6 Sol?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 42.3 on the Noometry Index. DeepSeek-R1 costs 8.8× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 or GPT-5.6 Sol?
DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; GPT-5.6 Sol lists at $4 and $20.
Is DeepSeek-R1 or GPT-5.6 Sol better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 46.3 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Sol does, with 1.05M tokens against 164K.
How many benchmarks do DeepSeek-R1 and GPT-5.6 Sol share?
31 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and GPT-5.6 Sol has 65.