DeepSeek-V4-Flash vs MiMo V2.5
Wins 7 of 7 areas
Coding · Agents · Reasoning · Facts · Math · Long documents · Following instructions
Wins 0 of 7 areas
—
DeepSeek-V4-Flash is the stronger all-rounder.MiMo V2.5 is cheaper.
Scores updated · 39 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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.
- AgentsCarrying out multi-step tasks on its own30DeepSeek-V4-Flash3 of 3 tests
- ReasoningHard problems that need careful thinking30DeepSeek-V4-Flash3 of 3 tests
- CodingWriting and fixing software42DeepSeek-V4-Flash4 of 7 tests · 1 tie
- MathCompetition and research-level math20DeepSeek-V4-Flash2 of 2 tests
- FactsGetting facts right instead of making them up21DeepSeek-V4-Flash2 of 3 tests
- Long documentsFinding answers in very long texts10DeepSeek-V4-Flash1 of 1 test
- Following instructionsDoing exactly what it is asked10DeepSeek-V4-Flash1 of 1 test
Long documents and following instructions 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.
MiMo V2.5 costs 18% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where DeepSeek-V4-Flash pulls ahead
- Recent programming contest problemsLiveCodeBench v6+55.1points ahead
- Fixes real GitHub issues in Python projectsSWE-bench Verified+48.2points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+30.7points ahead
Where MiMo V2.5 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+59.8points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+6.1points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+3.5points ahead
Every test, side by side
All 39 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingDeepSeek-V4-Flash
- LiveCodeBench v6DeepSeek-V4-Flash by 55.190.635.5+55.1
- SWE-bench VerifiedDeepSeek-V4-Flash by 48.27930.8+48.2
- Terminal-Bench 2.1DeepSeek-V4-Flash by 1578.763.7+15
- SciCodeDeepSeek-V4-Flash by 6.450.343.9+6.4
- Terminal-Bench HardMiMo V2.5 by 6.135.641.7+6.1
- SWE-bench ProMiMo V2.5 by 3.552.656.1+3.5
- LMArena · WebDevtie14311438tie
AgentsDeepSeek-V4-Flash
- τ-Bench V3 · BankingDeepSeek-V4-Flash by 30.739.48.7+30.7
- GDPValDeepSeek-V4-Flash by 21.846.724.9+21.8
- Terminal-Bench 4.0DeepSeek-V4-Flash by 12.112.10+12.1
ReasoningDeepSeek-V4-Flash
- CritPtDeepSeek-V4-Flash by 12.916.63.7+12.9
- Humanity's Last ExamDeepSeek-V4-Flash by 11.438.627.2+11.4
- GPQA DiamondDeepSeek-V4-Flash by 5.990.884.9+5.9
FactsDeepSeek-V4-Flash
- AA-Omniscience · Non-hallucinationMiMo V2.5 by 59.88.368.1+59.8
- AA-Omniscience · AccuracyDeepSeek-V4-Flash by 23.640.416.8+23.6
- SimpleQA VerifiedDeepSeek-V4-Flash by 1834.116.1+18
MathDeepSeek-V4-Flash
- HMMT Feb. 2026DeepSeek-V4-Flash by 12.294.882.6+12.2
- AIME 2026DeepSeek-V4-Flash by 2.295.893.6+2.2
Following instructionsDeepSeek-V4-Flash
- IFBenchDeepSeek-V4-Flash by 12.179.267.1+12.1
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.
- Toolathlon VerifiedDeepSeek-V4-Flash by 21.270.349.1+21.2
- MMLU-ProDeepSeek-V4-Flash by 20.486.265.8+20.4
- τ³-Bench BankingDeepSeek-V4-Flash by 13.922.99+13.9
- AA IntelligenceDeepSeek-V4-Flash by 12.534.321.8+12.5
- MATHMiMo V2.5 by 10.357.467.7+10.3
- Terminal-Bench 2.0MiMo V2.5 by 8.956.965.8+8.9
- GDPval-AA v2MiMo V2.5 by 76 rating points11891265+76 rating
- WinoGrandeMiMo V2.5 by 5.279.584.7+5.2
- HLE (with tools)DeepSeek-V4-Flash by 5.145.140+5.1
- DROPDeepSeek-V4-Flash by 4.988.683.7+4.9
- AA-OmniscienceMiMo V2.5 by 4.5-14.3-9.8+4.5
- τ²-Bench Telecom (AA run)DeepSeek-V4-Flash by 4.49590.6+4.4
- SWEBench Pro PublicMiMo V2.5 by 3.552.656.1+3.5
- HellaSwagMiMo V2.5 by 2.985.788.6+2.9
- MMLUDeepSeek-V4-Flash by 2.488.786.3+2.4
- CMMLUDeepSeek-V4-Flash by 2.290.488.2+2.2
- TriviaQADeepSeek-V4-Flash by 2.182.880.7+2.1
- StrongREJECTMiMo V2.5 by 1.997.499.3+1.9
- MMLU-Reduxtie89.489.8tie
Questions people ask
Which is better, DeepSeek-V4-Flash or MiMo V2.5?
DeepSeek-V4-Flash wins all seven areas where both have results: coding, agents, reasoning, facts, math, long documents and following instructions. MiMo V2.5 wins none, but costs 18% less.
Which is better for coding?
DeepSeek-V4-Flash. It wins 4 of the 7 coding tests both models report; MiMo V2.5 wins 2, and 1 is a tie.
Which is cheaper?
DeepSeek-V4-Flash costs $0.14 per million input tokens and $0.28 per million output tokens; MiMo V2.5 costs $0.11 and $0.22. That makes MiMo V2.5 about 18% cheaper for the same work.
How do you compare the two?
We use the 39 benchmark tests both models have published scores on. The verdict counts the 20 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.