Inference
EN

DeepSeek V3 0324 vs Ling 2.6 Flash

Change models
DeepSeek V3 0324DeepSeek
Ling 2.6 FlashInclusionAI
IndexAll categories, out of 100.
37.1 ±5.0 best
36.9 ±7.9
GapIndex points below the top.
Top
−0.2 · no clear gap
CoverageIndex weight that has results.
90% best
75%
SpeedOutput tokens per second.
—
104/s
EloLMArena rating and rank.
1375
—
ReleasedMonth of release.
2025-03
2026-04
Input / 1MUS dollars per 1M input tokens.
—
—
Output / 1MUS dollars per 1M output tokens.
—
—
1K chats2K tokens in, 500 out, each.
—
—
1K documents20K tokens in, 1K out, each.
—
—
1K agent tasks200K tokens in, 20K out, each.
—
—
ContextMaximum tokens in one request.
128K
262K best
AgenticMulti-step tasks with tools.
34.1
46.9 best
CodingCode writing and repair.
40.8 best
31.0
ReasoningLogic problems and puzzles.
36.1 best
32.9
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
37.6 best
30.8
MultilingualTasks in many languages.
40.0
—
InstructionTasks with strict prompt rules.
27.5
42.5 best
MathMath problems.
38.0
—

Benchmarks

8 shared
Only DeepSeek V3 0324: 14Only Ling 2.6 Flash: 3
DeepSeek V3 0324
Ling 2.6 Flash
Artificial Analysis Humanity's Last ExamKnowledge
4.7%183 / 222
6.3% best168 / 222
Artificial Analysis Long Context ReasoningReasoning
40.7% best167 / 217
31.3%179 / 217
Critical Physics TasksReasoning
0.0%150 / 214
0.0%150 / 214
Artificial Analysis GPQA DiamondKnowledge
65.5% best144 / 199
59.3%157 / 199
Artificial Analysis Omniscience AccuracyKnowledge
24.3% best109 / 165
15.6%150 / 165
τ²-Bench Tool-Agent-User EvaluationAgentic
47.1%97 / 149
86.0% best54 / 149
Artificial Analysis IFBenchInstruction
41.0%102 / 146
57.4% best67 / 146
GDPval-AA normalizedAgentic
0.0%101 / 133
0.0%101 / 133