DeepSeek-R1 vs o3
Wins 0 of 6 areas
—
Wins 6 of 6 areas
Coding · Agents · Reasoning · Facts · Long documents · Following instructions
o3 is the stronger all-rounder.DeepSeek-R1 is cheaper.
Scores updated · 36 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.
- CodingWriting and fixing software03o33 of 3 tests
- ReasoningHard problems that need careful thinking02o32 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up02o32 of 2 tests
- Following instructionsDoing exactly what it is asked02o32 of 2 tests
- Long documentsFinding answers in very long texts01o31 of 2 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own01o31 of 1 test
Agents 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-R1 costs 40% 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 o3 pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+32.4points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+31points ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+19.9points ahead
Where DeepSeek-R1 pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 36 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Codingo3
- Terminal-Bench Hardo3 by 316.137.1+31
- SWE-bench Verifiedo3 by 19.949.269.1+19.9
- SciCodeo3 by 2.738.341+2.7
Reasoningo3
- GPQA Diamondo3 by 11.970.882.7+11.9
- Humanity's Last Examo3 by 11.68.520.1+11.6
- CritPttie0.61.1tie
Factso3
- AA-Omniscience · Accuracyo3 by 8.130.538.6+8.1
- AA-Omniscience · Non-hallucinationo3 by 2.29.711.9+2.2
Following instructionso3
- IFBencho3 by 32.43971.4+32.4
- Multi-Challengeo3 by 19.740.760.4+19.7
Other results23 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)o3 by 69.311.480.7+69.3
- HarmBencho3 by 50.547.998.4+50.5
- ARC-AGI-1o3 by 4515.860.8+45
- HMMT Feb. 2025o3 by 35.841.777.5+35.8
- AIR-Bench 2024o3 by 31.652.984.5+31.6
- Aider-Polygloto3 by 2853.381.3+28
- LiveCodeBencho3 by 21.263.584.7+21.2
- OpenAI-MRCR (128k)o3 by 20.735.856.5+20.7
- LiveCodeBench (24/8~25/5)o3 by 19.955.975.8+19.9
- SimpleQAo3 by 19.330.149.4+19.3
- ZebraLogico3 by 17.178.795.8+17.1
- AA-Omniscienceo3 by 16.7-32.2-15.6+16.7
- AIME 2025o3 by 16.47086.4+16.4
- Artificial Analysis Coding Indexo3 by 13.824.638.4+13.8
- AIME 2024o3 by 11.879.891.6+11.8
- AA Intelligenceo3 by 8.811.420.2+8.8
- XSTesto3 by 2.994.497.3+2.9
- bbqo3 by 1.396.697.9+1.3
- anthropic_red_teamo3 by 1.197.298.3+1.1
- MMLU-Proo3 by 18485+1
- simple_safety_testso3 by 19899+1
- FullStackBenchtie70.169.3tie
- MATH-500 (EM)tie97.398.1tie
Questions people ask
Which is better, DeepSeek-R1 or o3?
o3 wins all six areas where both have results: coding, agents, reasoning, facts, long documents and following instructions. DeepSeek-R1 wins none, but costs 40% less.
Which is better for coding?
o3. It wins 3 of the 3 coding tests both models report; DeepSeek-R1 wins none.
Which is cheaper?
DeepSeek-R1 costs $2.00 per million input tokens and $4.00 per million output tokens; o3 costs $2.00 and $8.00. That makes DeepSeek-R1 about 40% cheaper for the same work.
How do you compare the two?
We use the 36 benchmark tests both models have published scores on. The verdict counts the 13 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 23 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.