DeepSeek-V3 is the stronger all-rounder.
Scores updated · 24 tests both models report · How we compare
Where each one wins
Tests won in each of the three 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.
Coding and facts rest on a single test each.
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 costs 91% 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 pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+33.2points ahead
- Code for real scientific research problemsSciCode+27.2points ahead
- Common-sense trick questionsSimpleBench+1.5points ahead
Where Command R+ (Apr '24) pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 24 tests both models report. The winning score is in its model's colour; marks a score checked independently.
ReasoningDeepSeek-V3
- GPQA DiamondDeepSeek-V3 by 33.232.365.5+33.2
- SimpleBenchDeepSeek-V3 by 1.517.418.9+1.5
- Humanity's Last Examtie4.64.7tie
Other results19 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.
- MATHDeepSeek-V3 by 49.940.390.2+49.9
- MMLUDeepSeek-V3 by 29.55988.5+29.5
- ARC-ChallengeDeepSeek-V3 by 24.37195.3+24.3
- GSM8KDeepSeek-V3 by 23.370.794+23.3
- OpenBookQADeepSeek-V3 by 12.682.895.4+12.6
- naturalquestions_closedbookDeepSeek-V3 by 12.434.346.7+12.4
- AIR-Bench 2024DeepSeek-V3 by 11.529.340.8+11.5
- vectara_avg_summary_lengthCommand R+ (Apr '24) by 9.891.581.7+9.8
- bbqDeepSeek-V3 by 6.889.996.7+6.8
- NarrativeQADeepSeek-V3 by 6.173.579.6+6.1
- simple_safety_testsCommand R+ (Apr '24) by 4.710095.3+4.7
- AA IntelligenceDeepSeek-V3 by 4.45.39.7+4.4
- XSTestDeepSeek-V3 by 3.393.897.1+3.3
- vectara_answer_rateDeepSeek-V3 by 2.59597.5+2.5
- HarmBenchDeepSeek-V3 by 1.248.549.7+1.2
- anthropic_red_teamtie9897.1tie
- vectara_factual_consistencytie93.193.9tie
- WinoGrandetie85.484.9tie
- HellaSwagtie88.688.9tie
Questions people ask
Which is better, Command R+ (Apr '24) or DeepSeek-V3?
DeepSeek-V3 wins two of the three areas where both have results: coding and reasoning. Command R+ (Apr '24) wins none. They are level on facts.
Which is better for coding?
DeepSeek-V3. It wins the one coding test both models report.
Which is cheaper?
Command R+ (Apr '24) costs $2.50 per million input tokens and $10.00 per million output tokens; DeepSeek-V3 costs $0.24 and $0.90. That makes DeepSeek-V3 about 91% cheaper for the same work.
How do you compare the two?
We use the 24 benchmark tests both models have published scores on. The verdict counts the 5 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 19 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.