Compare / DeepSeek vs OpenAI
DeepSeek vs OpenAI: which AI lab is ahead?
Each lab's best model on sale goes head to head on every kind of work we measure, from our own benchmarks and licensed sources. 39 models in all; a lab is only as good as the model you would actually pick from it.
Kinds of work led by each lab's best model there, of the 8 both are measured on.
Results as of 8 October 2026 (catalogue release 2026-10-08-4199ce0620fe). Models on sale means listed on OpenRouter today.
DeepSeek or OpenAI? The short answer
OpenAI is ahead: its best models beat DeepSeek's in all 8 kinds of work both are measured on.
- Pick OpenAI for product listings, decks from an analysis, marketing planning, agents and tool use, professional work, reasoning and knowledge, writing and sticking to the facts.
- On our overall leaderboard, GPT-6.1 Sol is #2 (81 of 100) and DeepSeek-V4.1-Flash is #9 (56).
- Prices start at $0.18 per million tokens for DeepSeek (DeepSeek-V4-Flash (0731)) and $0.060 for OpenAI (gpt-oss-20b).
Head to head
Each line names the model behind the number. Prices are blended, three input tokens to one output, at the developer's list price where we have read it; see every model's price. Intelligence Index from Artificial Analysis; overall score from our leaderboard.
Which lab is better for what
For each kind of work, each lab's best model there and its rank among every model on sale. Further right is better; the better rank is in green.
Every model, by intelligence and price
Each dot is a model on sale with an Intelligence Index from Artificial Analysis. Up is smarter; left is cheaper. Hover or tap a dot for its name.
DeepSeekOpenAI
On our overall leaderboard
The 6 of their models we rank overall: one score out of 100 across everything we publish (how the score works). Only models with enough of our own results are ranked.
- #2GPT-6.1 Sol81
- #3GPT-6 Astra79
- #4GPT-6 Sol69
- #9DeepSeek-V4.1-Flash56
- #15GPT-6 Luna40
- #18DeepSeek-V4-Pro (0423)33
Benchmark by benchmark
The headline result of every benchmark both labs have models on: each lab's best model and its rank among the models on sale. Green marks the better rank.
| Measure | DeepSeek | OpenAI |
|---|---|---|
| Agents and tool use | ||
| Berkeley Function Calling Leaderboard (BFCL) V4 · Overall accuracy · BFCL% correct, higher is better | 56.7%DeepSeek-V3.2-Exp · 7 of 38 | 63.1%o3 · 5 of 38 |
| OpenHands Index · Average% resolved, higher is better | 45.7%DeepSeek-V3.2 · 20 of 33 | 65.9%GPT-5.5 · 5 of 33 |
| Microsoft STATE-Bench · Task success (pass@1) · STATE-Bench v0.7% of tasks, higher is better | 47.2%DeepSeek-V4-Pro (0423) · 4 of 4 | 55.7%GPT-5.4 · 1 of 4 |
| Artificial Analysis · Terminal-Bench 4.0% of tasks, higher is better | 26.8%DeepSeek-V4.1-Flash · 17 of 105 | 59.6%GPT-6 Astra · 2 of 105 |
| Artificial Analysis · Terminal-Bench 2.1% of tasks, higher is better | 78.7%DeepSeek-V4-Pro (0813) · 24 of 115 | 89.9%GPT-6 Astra · 2 of 115 |
| Artificial Analysis · τ-bench banking% of tasks, higher is better | 39.6%DeepSeek-V4-Pro (0813) · 17 of 113 | 44.3%GPT-5.6 Sol · 12 of 113 |
| Vending-Bench 2 · Money after a yearUS dollars, higher is better | $3,285DeepSeek-V4-Pro (0423) · 37 of 56 | $15,515GPT-6 Astra · 1 of 56 |
| Decks from an analysis | ||
| DeckBench · Deck ratingrating, higher is better | 613DeepSeek-V4-Pro (0423) · 18 of 19 | 1,477GPT-6 Astra · 1 of 19 |
| Marketing planning | ||
| ROASBench · Overall scorescore out of 100, higher is better | 28.3DeepSeek-V4-Pro (0423) · 13 of 19 | 56.6GPT-6 Astra · 1 of 19 |
| Product listings | ||
| CatalogBench · Reliably publish-ready% of products, higher is better | 41.1%DeepSeek-V4.1-Flash · 9 of 20 | 75.0%GPT-6.1 Sol · 1 of 20 |
| CatalogBench · Reliably publish-ready · sales brief% of products, higher is better | 23.2%DeepSeek-V4.1-Flash · 6 of 20 | 69.6%GPT-6 Astra · 1 of 20 |
| Professional work | ||
| APEX-Agents · Tasks passed% of tasks, higher is better | 47.3%DeepSeek-V4-Pro (0813) · 24 of 37 | 64.7%GPT-6 Astra · 7 of 37 |
| Reasoning and knowledge | ||
| Artificial Analysis · Artificial Analysis Intelligence Indexindex score, higher is better | 39.5DeepSeek-V4.1-Flash · 26 of 198 | 52.7GPT-6 Astra · 4 of 198 |
| Artificial Analysis · Humanity's Last Exam% of questions, higher is better | 41.0%DeepSeek-V4-Pro (0813) · 34 of 197 | 54.7%GPT-6 Astra · 6 of 197 |
| Artificial Analysis · GPQA Diamond% of questions, higher is better | 92.8%DeepSeek-V4-Pro (0813) · 15 of 187 | 96.3%GPT-6 Astra · 1 of 187 |
| Sticking to the facts | ||
| Vectara Hallucination Leaderboard · Hallucination rate · Vectara% of summaries, lower is better | 5.3%DeepSeek-V3.2-Exp · 10 of 79 | 3.1%GPT-5.4 nano · 1 of 79 |
| Arena (formerly LMArena) · Overall · Text factualityArena rating, higher is better | 1,467DeepSeek-V4.1-Flash · 23 of 113 | 1,486GPT-5.5 · 5 of 113 |
| SimpleQA Verified (Epoch AI) · Correct answers% of questions, higher is better | 52.9%DeepSeek-V4-Pro (0813) · 20 of 73 | 75.6%GPT-6 Astra · 1 of 73 |
| Writing | ||
| Arena (formerly LMArena) · Overall · TextArena rating, higher is better | 1,474DeepSeek-V4.1-Flash · 26 of 166 | 1,484GPT-5.6 Sol · 14 of 166 |
| UGI Leaderboard · Writing score · UGIscore out of 100, higher is better | 68.4DeepSeek-V4-Pro (0423) · 14 of 138 | 69.7GPT-5.5 · 11 of 138 |
Their models, one against one
- DeepSeek-V4.1-Flash vs GPT-6.1 SolBenchmarks, price and which to pick
- DeepSeek-V4.1-Flash vs GPT-6 AstraBenchmarks, price and which to pick
- DeepSeek-V4.1-Flash vs GPT-6 SolBenchmarks, price and which to pick
- DeepSeek-V4.1-Flash vs GPT-6 LunaBenchmarks, price and which to pick
- DeepSeek-V4-Pro (0423) vs GPT-6.1 SolBenchmarks, price and which to pick
- DeepSeek-V4-Pro (0423) vs GPT-6 AstraBenchmarks, price and which to pick
- DeepSeek-V4-Pro (0423) vs GPT-6 SolBenchmarks, price and which to pick
- DeepSeek-V4-Pro (0423) vs GPT-6 LunaBenchmarks, price and which to pick
Quick answers
Is DeepSeek or OpenAI better?
OpenAI is ahead: its best models beat DeepSeek's in all 8 kinds of work both are measured on.
What is DeepSeek's best model?
DeepSeek-V4.1-Flash: #9 on the Spring Prompt overall leaderboard and 39.5 on the Artificial Analysis Intelligence Index.
What is OpenAI's best model?
GPT-6.1 Sol: #2 on the Spring Prompt overall leaderboard and 51.8 on the Artificial Analysis Intelligence Index.
Which is better for agents and tool use, DeepSeek or OpenAI?
OpenAI: its best model there, GPT-6.1 Sol, ranks 3 of 164 models on sale, against 21 for DeepSeek's DeepSeek-V4.1-Flash.
Which is better for reasoning and knowledge, DeepSeek or OpenAI?
OpenAI: its best model there, GPT-6 Astra, ranks 3 of 198 models on sale, against 29 for DeepSeek's DeepSeek-V4-Pro (0813).
Which is better for writing, DeepSeek or OpenAI?
OpenAI: its best model there, GPT-5.5, ranks 14 of 188 models on sale, against 28 for DeepSeek's DeepSeek-V4.1-Flash.
Which is cheaper, DeepSeek or OpenAI?
OpenAI's cheapest model with results, gpt-oss-20b, costs $0.060 per million tokens (three input to one output), against $0.18 for DeepSeek-V4-Flash (0731) from DeepSeek.
Other labs compared
- Anthropic vs OpenAIBest model against best model
- Google vs OpenAIBest model against best model
- OpenAI vs xAIBest model against best model
- Meta vs OpenAIBest model against best model
- Anthropic vs DeepSeekBest model against best model
- DeepSeek vs GoogleBest model against best model
- DeepSeek vs xAIBest model against best model
- DeepSeek vs MetaBest model against best model
Which model is best for your product?
Benchmarks aren't your data. We test the models from DeepSeek, OpenAI and the rest on your own tasks, with the cost of every answer.