DeepSeek-V4-Flash vs Inkling-Small
Wins 5 of 7 areas
Agents · Reasoning · Facts · Math · Long documents
Wins 1 of 7 areas
Following instructions
DeepSeek-V4-Flash is the stronger all-rounder.Inkling-Small is cheaper and better at following instructions.
Scores updated · 36 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.
- ReasoningHard problems that need careful thinking40DeepSeek-V4-Flash4 of 4 tests
- MathCompetition and research-level math30DeepSeek-V4-Flash3 of 4 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own32DeepSeek-V4-Flash3 of 5 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 asked01Inkling-Small1 of 1 test
- CodingWriting and fixing software22Even2 each · 1 tie
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.
Inkling-Small costs 15% 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
- Command-line tasks in a real terminalTerminal-Bench 2.1+23.6points ahead
- Abstract visual puzzles that people can solveARC-AGI-2+21.3points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+20.6points ahead
Where Inkling-Small pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+28.7points ahead
- Multi-step tasks using many real software toolsMCP Atlas+10.6points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+4.2points ahead
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.
CodingEven
- Terminal-Bench 2.1DeepSeek-V4-Flash by 23.678.755.1+23.6
- SWE-bench ProInkling-Small by 3.352.655.9+3.3
- LMArena · WebDevDeepSeek-V4-Flash by 21 rating points14301409+21 rating
- SWE-bench VerifiedInkling-Small by 1.27980.2+1.2
- SciCodetie50.349.7tie
AgentsDeepSeek-V4-Flash
- τ-Bench V3 · BankingDeepSeek-V4-Flash by 20.639.418.8+20.6
- GDPValDeepSeek-V4-Flash by 15.746.931.2+15.7
- Terminal-Bench 4.0DeepSeek-V4-Flash by 11.112.11+11.1
- MCP AtlasInkling-Small by 10.66979.6+10.6
- BrowseCompInkling-Small by 4.273.277.4+4.2
ReasoningDeepSeek-V4-Flash
- ARC-AGI-2DeepSeek-V4-Flash by 21.361.440.1+21.3
- CritPtDeepSeek-V4-Flash by 8.316.68.3+8.3
- Humanity's Last ExamDeepSeek-V4-Flash by 5.338.633.3+5.3
- GPQA DiamondDeepSeek-V4-Flash by 1.390.889.5+1.3
FactsDeepSeek-V4-Flash
- AA-Omniscience · Non-hallucinationInkling-Small by 28.78.337+28.7
- SimpleQA VerifiedDeepSeek-V4-Flash by 13.534.120.6+13.5
- AA-Omniscience · AccuracyDeepSeek-V4-Flash by 7.240.433.2+7.2
MathDeepSeek-V4-Flash
- FrontierMath Tiers 1-3 (v2)DeepSeek-V4-Flash by 11.257.546.3+11.2
- FrontierMath Tier 4DeepSeek-V4-Flash by 7.324.417.1+7.3
- HMMT Feb. 2026DeepSeek-V4-Flash by 4.694.890.2+4.6
- AIME 2026tie95.895.5tie
Following instructionsInkling-Small
- IFBenchInkling-Small by 379.282.2+3
Other results13 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.
- AA Agentic IndexDeepSeek-V4-Flash by 16.741.725+16.7
- Artificial Analysis Coding IndexDeepSeek-V4-Flash by 16.269.152.9+16.2
- ToolathlonDeepSeek-V4-Flash by 15.970.354.4+15.9
- Toolathlon VerifiedDeepSeek-V4-Flash by 15.970.354.4+15.9
- AA IntelligenceDeepSeek-V4-Flash by 8.634.325.7+8.6
- τ³-Bench BankingDeepSeek-V4-Flash by 7.422.915.5+7.4
- AA-OmniscienceInkling-Small by 5.4-14.3-8.9+5.4
- ARC-AGI-1DeepSeek-V4-Flash by 58984+5
- GDPval-AA v2Inkling-Small by 49 rating points11891238+49 rating
- BrowseComp (context management)Inkling-Small by 4.273.277.4+4.2
- SWEBench Pro PublicInkling-Small by 3.352.655.9+3.3
- HLE (with tools)Inkling-Small by 2.745.147.8+2.7
- StrongREJECTInkling-Small by 197.498.4+1
Questions people ask
Which is better, DeepSeek-V4-Flash or Inkling-Small?
DeepSeek-V4-Flash wins five of the seven areas where both have results: agents, reasoning, facts, math and long documents. Inkling-Small wins following instructions, and costs 15% less. They are level on coding.
Which is better for coding?
Neither. They win 2 coding tests each of the 5 both models report, and 1 is a tie.
Which is cheaper?
DeepSeek-V4-Flash costs $0.44 per million input tokens and $1.32 per million output tokens; Inkling-Small costs $0.30 and $1.20. That makes Inkling-Small about 15% 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 23 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 13 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.