DeepSeek-R1 vs DeepSeek-V3.1-Terminus
Wins 0 of 6 areas
—
Wins 5 of 6 areas
Coding · Agents · Reasoning · Long documents · Following instructions
DeepSeek-V3.1-Terminus is the stronger all-rounder.
Scores updated · 19 tests both models report · How we compare
Where each one wins
Tests won in each of the six areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- ReasoningHard problems that need careful thinking03DeepSeek-V3.1-Terminus3 of 3 tests
- CodingWriting and fixing software02DeepSeek-V3.1-Terminus2 of 3 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own02DeepSeek-V3.1-Terminus2 of 3 tests · 1 tie
- Following instructionsDoing exactly what it is asked02DeepSeek-V3.1-Terminus2 of 2 tests
- Long documentsFinding answers in very long texts01DeepSeek-V3.1-Terminus1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
Long documents rests on a single test.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
DeepSeek-V3.1-Terminus costs 79% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where DeepSeek-V3.1-Terminus pulls ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+25.8points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+24.2points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+18points ahead
Where DeepSeek-R1 pulls ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+2.8points ahead
Every test, side by side
All 19 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-V3.1-Terminus
- Terminal-Bench 2.1DeepSeek-V3.1-Terminus by 25.819.144.9+25.8
- Terminal-Bench HardDeepSeek-V3.1-Terminus by 24.26.130.3+24.2
- SciCodetie38.338tie
AgentsDeepSeek-V3.1-Terminus
- τ-Bench V3 · BankingDeepSeek-V3.1-Terminus by 14.66.421+14.6
- GDPValDeepSeek-V3.1-Terminus by 10.6010.6+10.6
- Terminal-Bench 4.0tie00tie
ReasoningDeepSeek-V3.1-Terminus
- GPQA DiamondDeepSeek-V3.1-Terminus by 8.470.879.2+8.4
- Humanity's Last ExamDeepSeek-V3.1-Terminus by 7.98.516.4+7.9
- CritPtDeepSeek-V3.1-Terminus by 1.10.61.7+1.1
FactsEven
- AA-Omniscience · Non-hallucinationDeepSeek-V3.1-Terminus by 15.59.725.2+15.5
- AA-Omniscience · AccuracyDeepSeek-R1 by 2.830.527.7+2.8
Long documentsDeepSeek-V3.1-Terminus
- AA-LCRDeepSeek-V3.1-Terminus by 11.657.769.3+11.6
Following instructionsDeepSeek-V3.1-Terminus
- IFBenchDeepSeek-V3.1-Terminus by 183957+18
- Multi-ChallengeDeepSeek-V3.1-Terminus by 13.740.754.4+13.7
Other results5 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- τ²-Bench Telecom (AA run)DeepSeek-V3.1-Terminus by 25.711.437.1+25.7
- Artificial Analysis Coding IndexDeepSeek-V3.1-Terminus by 18.924.643.5+18.9
- LiveCodeBenchDeepSeek-V3.1-Terminus by 11.463.574.9+11.4
- AA-OmniscienceDeepSeek-V3.1-Terminus by 5.8-32.2-26.4+5.8
- AA IntelligenceDeepSeek-V3.1-Terminus by 3.411.414.8+3.4
Questions people ask
Which is better, DeepSeek-R1 or DeepSeek-V3.1-Terminus?
DeepSeek-V3.1-Terminus wins five of the six areas where both have results: coding, agents, reasoning, long documents and following instructions. DeepSeek-R1 wins none. They are level on facts.
Which is better for coding?
DeepSeek-V3.1-Terminus. It wins 2 of the 3 coding tests both models report; DeepSeek-R1 wins none, and 1 is a tie.
Which is cheaper?
DeepSeek-R1 costs $2.00 per million input tokens and $4.00 per million output tokens; DeepSeek-V3.1-Terminus costs $0.27 and $1.00. That makes DeepSeek-V3.1-Terminus about 79% cheaper for the same work.
How do you compare the two?
We use the 19 benchmark tests both models have published scores on. The verdict counts the 14 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 5 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.