Inkling-Small vs Qwen3.5 397B A17B
Wins 6 of 8 areas
Coding · Agents · Reasoning · Facts · Math · Following instructions
Wins 2 of 8 areas
Images and charts · Long documents
Inkling-Small is the stronger all-rounder.Qwen3.5 397B A17B is better at images and charts.
Scores updated · 34 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software40Inkling-Small4 of 5 tests · 1 tie
- AgentsCarrying out multi-step tasks on its own40Inkling-Small4 of 4 tests
- MathCompetition and research-level math30Inkling-Small3 of 3 tests
- ReasoningHard problems that need careful thinking20Inkling-Small2 of 3 tests · 1 tie
- FactsGetting facts right instead of making them up21Inkling-Small2 of 3 tests
- Following instructionsDoing exactly what it is asked10Inkling-Small1 of 1 test
- Images and chartsUnderstanding pictures, charts and video02Qwen3.5 397B A17B2 of 2 tests
- Long documentsFinding answers in very long texts01Qwen3.5 397B A17B1 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.
Inkling-Small costs 64% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Inkling-Small pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+25.9points ahead
- Real work tasks from 44 professionsGDPVal+16.4points ahead
- Unpublished advanced math problemsFrontierMath Tiers 1-3 (v2)+15.1points ahead
Where Qwen3.5 397B A17B pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+5.4points ahead
- Harder college exam questions with imagesMMMU-Pro+3.3points ahead
- Reasons across sets of long documentsAA-LCR+1.6points ahead
Every test, side by side
All 34 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingInkling-Small
- SWE-bench ProInkling-Small by 555.950.9+5
- SciCodeInkling-Small by 4.949.744.8+4.9
- SWE-bench VerifiedInkling-Small by 3.880.276.4+3.8
- Terminal-Bench 2.1Inkling-Small by 3.855.151.3+3.8
- LMArena · WebDevtie14091400tie
AgentsInkling-Small
- GDPValInkling-Small by 16.431.214.8+16.4
- BrowseCompInkling-Small by 8.477.469+8.4
- τ-Bench V3 · BankingInkling-Small by 5.418.813.4+5.4
- Terminal-Bench 4.0Inkling-Small by 110+1
ReasoningInkling-Small
- CritPtInkling-Small by 6.68.31.7+6.6
- Humanity's Last ExamInkling-Small by 4.333.329+4.3
- GPQA Diamondtie89.589.3tie
FactsInkling-Small
- AA-Omniscience · Non-hallucinationInkling-Small by 25.93711.1+25.9
- SimpleQA VerifiedQwen3.5 397B A17B by 5.420.626+5.4
- AA-Omniscience · AccuracyInkling-Small by 2.433.230.8+2.4
Images and chartsQwen3.5 397B A17B
- MMMU-ProQwen3.5 397B A17B by 3.37477.3+3.3
- LMArena · VisionQwen3.5 397B A17B by 30 rating points12341264+30 rating
MathInkling-Small
- FrontierMath Tiers 1-3 (v2)Inkling-Small by 15.146.331.2+15.1
- AIME 2026Inkling-Small by 4.295.591.3+4.2
- HMMT Feb. 2026Inkling-Small by 2.390.287.9+2.3
Long documentsQwen3.5 397B A17B
- AA-LCRQwen3.5 397B A17B by 1.675.777.3+1.6
Following instructionsInkling-Small
- IFBenchInkling-Small by 3.482.278.8+3.4
Other results12 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.
- GDPval-AA v2Inkling-Small by 276 rating points1238962+276 rating
- AA-OmniscienceInkling-Small by 21.8-8.9-30.8+21.8
- ToolathlonInkling-Small by 16.154.438.3+16.1
- AA Agentic IndexInkling-Small by 14.42510.6+14.4
- Toolathlon VerifiedInkling-Small by 13.754.440.7+13.7
- AA IntelligenceInkling-Small by 7.325.718.4+7.3
- SWEBench Pro PublicInkling-Small by 555.950.9+5
- Artificial Analysis Coding IndexInkling-Small by 4.752.948.2+4.7
- τ³-Bench BankingInkling-Small by 2.115.513.4+2.1
- BrowseComp (context management)Qwen3.5 397B A17B by 1.277.478.6+1.2
- StrongREJECTQwen3.5 397B A17B by 198.499.4+1
- HLE (with tools)tie47.848.3tie
Questions people ask
Which is better, Inkling-Small or Qwen3.5 397B A17B?
Inkling-Small wins six of the eight areas we test: coding, agents, reasoning, facts, math and following instructions. Qwen3.5 397B A17B wins images and charts and long documents.
Which is better for coding?
Inkling-Small. It wins 4 of the 5 coding tests both models report; Qwen3.5 397B A17B wins none, and 1 is a tie.
Which is cheaper?
Inkling-Small costs $0.30 per million input tokens and $1.20 per million output tokens; Qwen3.5 397B A17B costs $0.60 and $3.60. That makes Inkling-Small about 64% cheaper for the same work.
How do you compare the two?
We use the 34 benchmark tests both models have published scores on. The verdict counts the 22 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 12 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.