Inkling vs Inkling-Small

Change models
InklingThinking Machines Lab
Inkling-SmallThinking Machines Lab
IndexAll categories, out of 100.
54.8 ±2.9
55.1 ±3.2 best
GapIndex points below the top.
−0.3 · no clear gap
Top
CoverageIndex weight that has results.
95%
95%
SpeedOutput tokens per second.
85/s best
80/s
EloLMArena rating and rank.
1444 best
1414
ReleasedMonth of release.
2026-07
2026-07
Input / 1MUS dollars per 1M input tokens.
$1.00
$0.45 best
Output / 1MUS dollars per 1M output tokens.
$4.05
$1.20 best
1K chats2K tokens in, 500 out, each.
$4.03
$1.50 best
1K documents20K tokens in, 1K out, each.
$24.05
$10.20 best
1K agent tasks200K tokens in, 20K out, each.
$281.00
$114.00 best
ContextMaximum tokens in one request.
524K
524K
AgenticMulti-step tasks with tools.
51.2
55.7 best
CodingCode writing and repair.
54.1
54.9 best
ReasoningLogic problems and puzzles.
56.1 best
55.7
MultimodalTasks with images and text.
55.0 best
54.7
KnowledgeFacts and expert knowledge.
57.1 best
52.0
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
66.0 best
65.8
MathMath problems.
54.1
55.3 best

Benchmarks

59 shared
Only Inkling: 19Only Inkling-Small: 5
Inkling
Inkling-Small
31.9%72 / 222
33.3% best70 / 222
77.3% best77 / 217
75.7%92 / 217
88.3%56 / 217
88.5% best53 / 217
5.4%81 / 214
8.3% best74 / 214
88.9%63 / 200
90.0% best62 / 200
87.2%59 / 199
89.5% best43 / 199
39.3% best56 / 164
31.8%74 / 164
82.9%64 / 149
83.0% best62 / 149
TaxEval v2Finance
75.3%19 / 143
75.5% best18 / 143
85.5%30 / 142
85.9% best25 / 142
87.9%39 / 141
89.5% best33 / 141
87.1% best40 / 135
83.6%55 / 135
MMLU ProKnowledge
86.3% best43 / 135
85.6%55 / 135
28.6%65 / 133
31.2% best60 / 133
36.5%54 / 131
40.1% best53 / 131
CorpFin v2Finance
68.6%7 / 131
69.6% best6 / 131
79.5%54 / 126
84.0% best52 / 126
19.2% best94 / 123
19.1%95 / 123
37.8%48 / 122
39.3% best46 / 122
90.6% best44 / 116
89.4%47 / 116