Kimi K3 vs Inkling 2 Small
Kimi K3 and Inkling 2 Small match in llmrun Score, with values of 132.8 versus 134.7; neither fits 24 GB.
Run Inkling 2 Small in the cloud →
Model A · Moonshot AI
Kimi K32779.9B · 105.8B active · 1M context · Other · Jun 2026
Model B · Thinking Machines
Inkling 2 Small594B · 17B active · 1M context · Apache 2.0 · Nov 2026
Choose Inkling 2 Small if you want…
Stronger at reasoning
136.1 vs 129.5
Lighter on memory
357.0 vs 1673.7 GB at Q4_K_M
More permissive license
Apache 2.0 vs Other
Newer release
Nov 2026 vs Jun 2026
Kimi K3: no clear edge over Inkling 2 Small in the data we have.
On your hardware
Each model at its default quant (Q4_K_M), a smaller one when it must. Estimates, ±20%.
8 GB
RTX 4060 · 8 GB tier
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
12 GB
RTX 3060 12GB · 12 GB tier
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
16 GB
RTX 5060 Ti 16GB · 16 GB tier
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
24 GBreference
RTX 4090 · 24 GB tier
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
32 GB
RTX 5090 · 32 GB tier
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
48 GB
RTX A6000 · 48 GB tier
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
Mac 32 GB
Mac Mini M4 · Macs
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
Mac 64 GB
Mac Studio M4 Max · Macs
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
Mac 128 GB
Mac Studio M4 Max · Macs
Kimi K3
Doesn't fit
smallest quant needs 766.5 GB
Inkling 2 Small
Doesn't fit
smallest quant needs 166.1 GB
Reference GPU: RTX 4090. Tinted cell = the better pick on that machine (fits, more headroom, or at least 10% faster).
Overall
Kimi K3
Inkling 2 Small
Coding
Kimi K3
Inkling 2 Small
no data
Agents
Kimi K3
Inkling 2 Small
no data
Math
Kimi K3
Inkling 2 Small
no data
Science
Kimi K3
Inkling 2 Small
no data
Reasoning
Kimi K3
Inkling 2 Small
Kimi K3 · #5 of 107 open models · #5 of 107 that fit ≤2 TB
Inkling 2 Small · #2 of 107 open models · #1 of 103 that fit ≤512 GB
Kimi K3
60.4%#7/23 open
IndependentEpoch AI · Jul 2026
Inkling 2 Small
81.7%#1/23 open
IndependentEpoch AI · Nov 2026
Kimi K3
59.5%#3/74 open
Lab-reportedEpoch AI · Jul 2026
Inkling 2 Small
52.3%#10/74 open
Lab-reportedEpoch AI · Nov 2026
Kimi K3
94.5%#4/23 open
IndependentEpoch AI · Jul 2026
Inkling 2 Small
96.5%#1/23 open
IndependentEpoch AI · Nov 2026
Kimi K3
47.8%#8/22 open
Lab-reportedEpoch AI · Jul 2026
Inkling 2 Small
46.8%#11/22 open
Lab-reportedEpoch AI · Nov 2026
39 more boards (other shared and single-model boards)
Kimi K3
1487.4#1/191 open
ArenaLMArena · Nov 2026
Inkling 2 Small
1411.8#46/191 open
ArenaLMArena · Nov 2026
Kimi K3
1657.8#3/58 open
ArenaLMArena · Nov 2026
Inkling 2 Small
1647.1#4/58 open
ArenaLMArena · Nov 2026
Only Kimi K3 (37)
- GPQA Diamond93.1%
- LiveBench Coding81.5
- LiveBench Reasoning90.7
- AIME 2024/202597.2%
- FrontierMath72.2%
- LiveBench Math84.4
- SimpleQA50.6%
- SimpleBench60.7%
- FrontierMath Tier 439.0%
- WeirdML82.6%
- DeepSWE68.5%
- FrontierCode44.2%
- Chess Puzzles37.0%
- CritPt23.4%
- DTBench91.2%
- ForecastBench64.2
- Furniture Assembly34.2%
- LiveBench 2026 Reasoning90.7
- LMCA52.7%
- Mystery Game Puzzles26.0%
- Vending-Bench 25165.0
- Blueprint-Bench 229.5%
- GDP.pdf19.0%
- LiveBench 2026 Browser16.0
- LiveBench 2026 Data Analysis78.7
- ALE-Bench1524.5
- FrontierSWE25.9%
- LiveBench 2026 Coding59.3
- LiveBench 2026 Long-Horizon42.8
- Surface Evolver Bench95.0%
- GBA Eval48.4%
- LiveBench 2026 Design77.2
- MirrorCode54.4%
- WeirdML v37.5%
- ProofBench87.0%
- FrontierMath Erdős0.0%
- LiveBench 2026 Research79.0
VRAM by quantization
Memory to load the model and 2K context; below it, the total at 32K and 128K context. Lighter side in bold.
Q2_K
Kimi K3
1187 GB928.6 GB file
32K 1185 GB · 128K 1196 GB
Inkling 2 Small
253 GB~252.4 GB file
32K 258 GB · 128K 272 GB
Q3_K_M
Kimi K3
1361 GB~1355.2 GB file
32K 1359 GB · 128K 1370 GB
Inkling 2 Small
290 GB~289.6 GB file
32K 295 GB · 128K 309 GB
Q4_K_M
Kimi K3
1674 GB~1668.0 GB file
32K 1672 GB · 128K 1682 GB
Inkling 2 Small
357 GB~356.4 GB file
32K 362 GB · 128K 376 GB
Q5_K_M
Kimi K3
1986 GB~1980.7 GB file
32K 1985 GB · 128K 1995 GB
Inkling 2 Small
424 GB~423.2 GB file
32K 428 GB · 128K 443 GB
Q6_K
Kimi K3
2299 GB~2293.4 GB file
32K 2297 GB · 128K 2308 GB
Inkling 2 Small
491 GB~490.1 GB file
32K 495 GB · 128K 510 GB
Q8_0
Kimi K3
2786 GB~2779.9 GB file
32K 2784 GB · 128K 2794 GB
Inkling 2 Small
595 GB~594.0 GB file
32K 599 GB · 128K 614 GB
3 more quants
IQ2_XXS
Kimi K3
770 GB711.1 GB file
32K 768 GB · 128K 779 GB
Inkling 2 Small
164 GB190.6 GB file
32K 168 GB · 128K 183 GB
IQ3_XXS
Kimi K3
not listed
Inkling 2 Small
231 GB226.0 GB file
32K 235 GB · 128K 250 GB
BF16
Kimi K3
5566 GB~5559.9 GB file
32K 5564 GB · 128K 5574 GB
Inkling 2 Small
1189 GB~1188.0 GB file
32K 1193 GB · 128K 1208 GB
Kimi K3
Ollama · Q4_K_M
ollama run hf.co/unsloth/Kimi-K3-GGUFInkling 2 Small
Ollama · Q4_K_M
ollama run hf.co/unsloth/Inkling-2-Small-GGUF| Spec | Kimi K3 | Inkling 2 Small |
|---|---|---|
| Publisher | Moonshot AI | Thinking Machines |
| Parameters | 2779.9B | 594B |
| Active parameters | 106B per token | 17B per token |
| Architecture | KimiK3ForConditionalGeneration | Inkling2ForConditionalGeneration |
| Context | 1M tokens | 1M tokens |
| Modalities | Vision, Chat | Vision, Chat |
| License | Other | Apache 2.0 |
| Released | Jun 2026 | Nov 2026 |
Frequently Asked Questions
Which is better, Kimi K3 or Inkling 2 Small?
Kimi K3 and Inkling 2 Small match in llmrun Score, with values of 132.8 versus 134.7; neither fits 24 GB. Choose Inkling 2 Small for stronger at reasoning and lighter on memory.
Which needs less VRAM, Kimi K3 or Inkling 2 Small?
At Q4_K_M, Kimi K3 requires 1673.7 GB while Inkling 2 Small needs 357.0 GB, making Inkling 2 Small the easier model to run on your local setup.
Can I run Kimi K3 or Inkling 2 Small on a 16 GB GPU?
Kimi K3 does not fit (smallest quant needs 766.5 GB); Inkling 2 Small does not fit (smallest quant needs 166.1 GB).
Which has a longer context window, Kimi K3 or Inkling 2 Small?
Kimi K3 supports 1,048,576 tokens and Inkling 2 Small supports 1,048,576 tokens.