Models / DeepSeek-V3.1

DeepSeek

DeepSeek-V3.1

DeepSeek-V3.1 is best for sticking to the facts; a strong pick at a modest price. The cheapest model of similar intelligence is gpt-oss-20b, at $0.06 per million tokens against $0.45.

Superseded. DeepSeek-V3.2 is DeepSeek's newer model in this line, at $0.32 blended: see DeepSeek-V3.2 or compare the two.

Results 16 Nov 2026 · 34 results · 4 sources

Test DeepSeek-V3.1 on your prompt → It runs beside the models you could use instead, with every answer checked and costed.

Price per million tokens
$0.27 in · $1.00 out
Typical host on OpenRouter, 10 Nov 2026 · compare prices
Speed
Not measured
Intelligence Index
13.7
Artificial Analysis, at no reasoning; 13.5–13.7 across 2 settings
Spring Prompt overall
Not ranked yet
needs our benchmarks and two groups of results
Developer
DeepSeek, based in China
Weights published
Released
21 Aug 2025
163,840 tokens of context

Where it stands among the models you could choose

Available models grouped by price, intelligence, speed and developer location, plus models with published weights. Retired models are excluded.

AmongIntelligence IndexSpring Prompt overallPrice, blendedSpeed
Models available today 116th of 185Claude Opus 5.5 – $0.45 blended, 35th percentilecheapest: Mistral Nemo –
Models priced $0.25–1 per million tokens 32nd of 61MiMo-V2.6-Pro – sets this group –
Models with similar intelligence sets this group – $0.45 blended, 48th percentilecheapest: gpt-oss-20b –
Models from developers based in China 50th of 75MiMo-V2.6-Pro – $0.45 blended, 41st percentilecheapest: Qwen3.5-9B –
Open-weights models 52nd of 104MiMo-V2.6-Pro – $0.45 blended, 51st percentilecheapest: Mistral Nemo –

The name under each rank is the leader of that group; green is the top quarter of the group and red the bottom quarter, by rank. Price is shown as a percentile: the share of the group that costs less, so lower is cheaper. "Sets this group" marks the measure the group is defined by. Spring Prompt overall is our score out of 100 across every source (how it works). Price is the developer's list price (or the typical OpenRouter host where we haven't read one) for three input tokens to one output; intelligence and speed are from Artificial Analysis (data sourced from Artificial Analysis); "similar intelligence" means within 4 points on its Intelligence Index. Developer locations are where each company is based, not where a model is served. Groups under 3 models are not ranked.

How good is it, and for what?

Each use case ranks the available models its benchmarks measured, then those priced $0.25–1 per million tokens. The bar shows its percentile in that field, best to the right. Open a row for the results behind it.

Product listingsTurning a sparse product feed and photos into listings that can go live Not measured
  • CatalogBench: has not measured this model.
Decks from an analysisTurning a finished analysis into a deck you could present as it is Not measured
  • DeckBench: has not measured this model.
User surveysPlanning a user survey and reading its results without being misled Not measured
  • SurveyBench: has not measured this model.
Marketing planningPlanning a year of ad spend without overspending Not measured
  • ROASBench: has not measured this model.
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls Not measured
  • tau2-bench: has not measured this model.
  • Berkeley Function Calling Leaderboard (BFCL) V4: has not measured this model.
  • OpenHands Index: has not measured this model.
  • Microsoft STATE-Bench: has not measured this model.
  • Artificial Analysis: has not measured this model.
  • Vending-Bench 2: has not measured this model.
Professional workReal tasks from banking, consulting and law, business documents and freelance projects Not measured
  • APEX-Agents: has not measured this model.
  • GDP.pdf: has not measured this model.
  • Remote Labor Index: has not measured this model.
Reasoning and knowledgeHard questions across science, maths and general knowledge 120th of 187 available · 34th of 61 at its price
BenchmarkDeepSeek-V3.1Rank among availableBest available
Artificial AnalysisArtificial Analysis: Artificial Analysis Intelligence Index · reported 13.7 116th of 185 Claude Opus 5.5 57.6
Artificial AnalysisArtificial Analysis: Humanity's Last Exam · reported 14.3% 117th of 186 Claude Opus 5.5 61.4%
Artificial AnalysisArtificial Analysis: GPQA Diamond · reported 77.9% 103rd of 168 GPT-6 Astra 96.3%
WritingWhat people prefer in blind comparisons, and judged writing quality 83rd of 169 available · 16th of 58 at its price
BenchmarkDeepSeek-V3.1Rank among availableBest available
Arena (formerly LMArena)Arena Text: overall · reported 1,417 83rd of 153 Gemini 4 Argon 1,525
UGI LeaderboardUGI: writing score · reported 51.9 55th of 124 Gemini 3.8 Flash 78.6
Sticking to the factsSummarising without inventing things, and factual answers 19th of 139 available · 2nd of 40 at its price
BenchmarkDeepSeek-V3.1Rank among availableBest available
Vectara Hallucination LeaderboardVectara: hallucination rate · reported 5.5% 11th of 71 GPT-5.4 nano 3.1%
  • Arena (formerly LMArena): has not measured this model.
  • SimpleQA Verified (Epoch AI): has not measured this model.
Speed under pressureGood decisions against a real clock (fast chess) Not measured
  • BulletBench: has not measured this model.

Nearest alternatives

Against DeepSeek-V3.1 at $0.45 blended, Intelligence Index 13.7. "Similar intelligence" means within 4 points or better.

Where to run it

4 hosts on OpenRouter at the standard service tier, with a median of $0.60 blended per million tokens; quantisation where the host states it: FP4, FP8. Cheapest first.

HostQuantisationInputOutputBlended
DeepInfraFP4$0.25$0.95$0.42
SiliconFlowFP8$0.27$1.00$0.45
CoreWeaveFP8$0.50$1.50$0.75
SambaNovaFP8$0.65$1.50$0.86

Prices per million tokens from OpenRouter's public endpoint list, checked 10 Nov 2026.

Against the alternatives

The models you would most likely weigh it against: the leaders of the groups above.

Scroll sideways to see every alternative.

DeepSeek-V3.1MiMo-V2.6-Pro
most intelligent at $0.25–1 per million tokens
gpt-oss-20b
cheapest with similar intelligence
Claude Opus 5.5
most intelligent available today
Price per million tokens $0.45 $0.54$0.06$8.00
Tokens a second – 4327796
Intelligence Index 13.7 46.310.057.6
Spring Prompt overall – ––82
Rank among available models, by use case
Product listings – ––13th
Decks from an analysis – ––7th
User surveys – ––4th
Marketing planning – ––9th
Agents and tool use – 23rd130th3rd
Professional work – ––3rd
Reasoning and knowledge 120th of 187 13th132nd1st
Writing 83rd of 169 23rd155th5th
Sticking to the facts 19th of 139 10th–3rd
Speed under pressure – –––

Shaded figures are better than DeepSeek-V3.1's. A dash means no result.

Where it has been measured

32nd most intelligent of the 61 models priced $0.25–1 per million tokens. Among models available today, in the top quarter (by percentile) for sticking to the facts.

Artificial Analysis lists it but has not published a speed for it yet

Every result

34 published results from 4 sources, each in the source's own units, with the configuration that produced it.

Arena (formerly LMArena) · reported by Arena (formerly LMArena) · 12 results
MeasureValueRankConfigurationDated
Arena Text: business, management and financeArena rating, higher is better 1,414
1,402–1,426
135 of 406 DeepSeek-V3.1 2 Oct 2026
Arena Text: business, management and financeArena rating, higher is better 1,417
1,404–1,431
129 of 406 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2026
Arena Text: creative writingArena rating, higher is better 1,388
1,375–1,402
131 of 411 DeepSeek-V3.1 2 Oct 2026
Arena Text: creative writingArena rating, higher is better 1,402
1,387–1,417
106 of 411 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2026
Arena Text: expert promptsArena rating, higher is better 1,426
1,404–1,447
148 of 363 DeepSeek-V3.1 2 Oct 2026
Arena Text: expert promptsArena rating, higher is better 1,432
1,407–1,457
140 of 363 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2026
Arena Text: instruction followingArena rating, higher is better 1,402
1,392–1,412
140 of 413 DeepSeek-V3.1 2 Oct 2026
Arena Text: instruction followingArena rating, higher is better 1,416
1,405–1,427
114 of 413 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2026
Arena Text: overallArena rating, higher is better 1,417
1,411–1,423
134 of 413 DeepSeek-V3.1 2 Oct 2026
Arena Text: overallArena rating, higher is better 1,416
1,409–1,423
136 of 413 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2026
Arena Text: writing, literature and languageArena rating, higher is better 1,398
1,388–1,408
131 of 412 DeepSeek-V3.1 2 Oct 2026
Arena Text: writing, literature and languageArena rating, higher is better 1,407
1,396–1,419
111 of 412 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2026

Not shown: Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Artificial Analysis · reported by Artificial Analysis · 14 results
MeasureValueRankConfigurationDated
Artificial Analysis: Artificial Analysis Intelligence Indexindex score, higher is better 13.7 270 of 413 DeepSeek-V3.1 (no reasoning) 20 Oct 2026
Artificial Analysis: Artificial Analysis Intelligence Indexindex score, higher is better 13.5 271 of 413 DeepSeek-V3.1 (thinking reasoning) 20 Oct 2026
Artificial Analysis: GPQA Diamond% of questions, higher is better 73.5% 229 of 360 DeepSeek-V3.1 (no reasoning) 20 Oct 2026
Artificial Analysis: GPQA Diamond% of questions, higher is better 77.9% 192 of 360 DeepSeek-V3.1 (thinking reasoning) 20 Oct 2026
Artificial Analysis: Humanity's Last Exam% of questions, higher is better 6.7% 320 of 419 DeepSeek-V3.1 (no reasoning) 20 Oct 2026
Artificial Analysis: Humanity's Last Exam% of questions, higher is better 14.3% 238 of 419 DeepSeek-V3.1 (thinking reasoning) 20 Oct 2026
Artificial Analysis: IFBench% of instructions, higher is better 37.8% 221 of 288 DeepSeek-V3.1 (no reasoning) 20 Oct 2026
Artificial Analysis: IFBench% of instructions, higher is better 41.5% 194 of 288 DeepSeek-V3.1 (thinking reasoning) 20 Oct 2026
Artificial Analysis: Long-context reasoning (AA-LCR)% of questions, higher is better 47.0% 287 of 407 DeepSeek-V3.1 (no reasoning) 20 Oct 2026
Artificial Analysis: Long-context reasoning (AA-LCR)% of questions, higher is better 56.7% 259 of 407 DeepSeek-V3.1 (thinking reasoning) 20 Oct 2026
Artificial Analysis: Terminal-Bench Hard% of tasks, higher is better 24.2% 132 of 282 DeepSeek-V3.1 (no reasoning) 20 Oct 2026
Artificial Analysis: Terminal-Bench Hard% of tasks, higher is better 25.0% 128 of 282 DeepSeek-V3.1 (thinking reasoning) 20 Oct 2026
Artificial Analysis: τ²-bench telecom% of tasks, higher is better 34.8% 192 of 286 DeepSeek-V3.1 (no reasoning) 20 Oct 2026
Artificial Analysis: τ²-bench telecom% of tasks, higher is better 37.4% 183 of 286 DeepSeek-V3.1 (thinking reasoning) 20 Oct 2026

Not shown: Not business work, and a blend: read the parts for any one task.

UGI Leaderboard · reported by UGI Leaderboard · 6 results
MeasureValueRankConfigurationDated
UGI: requested-length error% off the requested word count, lower is better 30.0% 262 of 379 DeepSeek-V3.1 (no reasoning) 2 Oct 2025
UGI: requested-length error% off the requested word count, lower is better 34.0% 278 of 379 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2025
UGI: style adherencescore from 0 to 1, higher is better 0.34 231 of 379 DeepSeek-V3.1 (no reasoning) 2 Oct 2025
UGI: style adherencescore from 0 to 1, higher is better 0.36 136 of 379 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2025
UGI: writing scorescore out of 100, higher is better 51.1 162 of 379 DeepSeek-V3.1 (no reasoning) 2 Oct 2025
UGI: writing scorescore out of 100, higher is better 51.9 156 of 379 DeepSeek-V3.1 (thinking reasoning) 2 Oct 2025

Not shown: Not other format limits such as character counts or bullet counts.

Vectara Hallucination Leaderboard · reported by Vectara Hallucination Leaderboard · 2 results
MeasureValueRankConfigurationDated
Vectara: answer rate% of documents, higher is better 94.5% 93 of 108 DeepSeek-V3.1 22 Sep 2026
Vectara: hallucination rate% of summaries, lower is better 5.5% 18 of 108 DeepSeek-V3.1 22 Sep 2026

Not shown: Not a quality score: a low rate usually means content filters were triggered, and hallucination rates are measured on answered documents only.

Will DeepSeek-V3.1 work for your product?

Benchmarks aren't your task. Paste the prompt your product uses and run it beside the models you could use instead, with every answer checked and costed.

New here? Your email starts a free trial. No credit card needed.