Inkling Small
| Released | 2026-07-30 |
|---|---|
| Price per 1M | $0.30 in · $1.20 out |
| Overall score | 52.7 · rank 94 |
| Context window | 1,048,576 tokens |
| Provider | Thinking Machines |
| Weights | Open weights (Apache-2.0) |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 52.7 | 94 | 31 |
| Overall: Coding | 52.5 | 86 | 21 |
| Overall: Reasoning | 54.8 | 80 | 12 |
| Overall: Agentic work | 53.4 | 89 | 19 |
| Overall: Frontend | 48.2 | 110 | 9 |
| Coding | 53.9 | 89 | 6 |
| Agents | 52.3 | 93 | 8 |
| Reasoning | 56.9 | 69 | 5 |
| Math | 54.9 | 71 | 4 |
| Knowledge | 49.8 | 93 | 3 |
| Frontend | 45.1 | 92 | 2 |
| General | 50.8 | 96 | 2 |
| Multimodal | 49.4 | 74 | 1 |
| Value vs price | +6.9 | 53 |
Configurations
| Configuration | Balanced | Benchmarks |
|---|---|---|
| Unspecified setting | 52.4 ±1.0 | 24 |
| High | 56.9 ±5.7 | 2 |
| Low | 50.2 ±5.7 | 2 |
| Medium | 53.5 ±5.7 | 2 |
| Minimal | 49.5 ±5.7 | 2 |
| Non-reasoning | 41.3 ±5.7 | 2 |
| Extra High | 53.1 ±4.4 | 8 |
Evidence
Coding 53.9 ±1.8
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Code Migration v1.1 | 10.4% | Vals AI | 2026-11-09 |
| IOI | 9.3% | Vals AI | 2026-10-29 |
| SWE-bench Verified (Vals run) | 82.2% | Vals AI | 2026-09-01 |
| SciCode | 49.7% | Epoch AI | 2026-11-11 |
| LiveCodeBench (Vals run) | 85.9% | Vals AI | 2026-09-01 |
| ProgramBench | 0.5% | Vals AI | 2026-11-09 |
Agents 52.3 ±2.0
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Terminal-Bench 4.0 (Vals run) | 0.0% | Vals AI | 2026-11-09 |
| Terminal-Bench 2.1 (Vals run) | 55.1% | Vals AI | 2026-10-01 |
| EMB v1.1 (Excel financial modeling) | 40.8% | Vals AI | 2026-11-09 |
| Tax Agent Bench | 15.4% | Vals AI | 2026-11-09 |
| Finance Agent v2.1 | 44.8% | Vals AI | 2026-11-09 |
| SkillsBench | 33.6% | Vals AI | 2026-10-29 |
| HLAB-Vals 1.1 (legal agent) | 2.9% | Vals AI | 2026-11-09 |
| CyberBench v1.1 | 64.1% | Vals AI | 2026-11-05 |
Reasoning 56.9 ±2.2
| Benchmark | Result | Source | Date |
|---|---|---|---|
| ARC-AGI-2 | 40.1% | Epoch AI | 2026-11-11 |
| ARC-AGI-2 (ARC Prize verified) | 40.1% | ARC Prize | 2026-11-11 |
| CritPt | 8.3% | Epoch AI | 2026-11-11 |
| ARC-AGI-1 (ARC Prize verified) | 84.0% | ARC Prize | 2026-11-11 |
| ARC-AGI-1 | 84.0% | Epoch AI | 2026-11-11 |
| GPQA Diamond | 88.5% | Epoch AI | 2026-08-14 |
| GPQA Diamond (Vals run) | 83.6% | Vals AI | 2026-09-01 |
| Mystery Game Puzzles | 6.0% | Epoch AI | 2026-08-15 |
Math 54.9 ±1.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| FrontierMath Tiers 1-3 | 46.3% | Epoch AI | 2026-08-14 |
| FrontierMath Tier 4 | 17.1% | Epoch AI | 2026-08-14 |
| OTIS Mock AIME 2024-2025 | 90.0% | Epoch AI | 2026-08-14 |
| ProofBench v1.1 | 6.0% | Vals AI | 2026-11-09 |
| ProofBench | 6.0% | Epoch AI | 2026-11-11 |
Knowledge 49.8 ±4.5
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SimpleQA Verified | 19.1% | Epoch AI | 2026-08-27 |
| MMLU-Pro (Vals run) | 85.6% | Vals AI | 2026-09-01 |
| Legal Research Bench | 25.5% | Vals AI | 2026-11-09 |
Frontend 45.1 ±0.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Vibe Code Bench v1.1 | 19.1% | Vals AI | 2026-11-09 |
| LMArena WebDev | 1,407 | LMArena | 2026-11-09 |
General 50.8 ±2.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| EQ-Bench Creative Writing v3 | 1,491 | EQ-Bench | 2026-11-11 |
| LMArena Text | 1,405 | LMArena | 2026-11-09 |
Multimodal 49.4 ±4.1
| Benchmark | Result | Source | Date |
|---|---|---|---|
| LMArena Vision | 1,208 | LMArena | 2026-11-09 |
Results not used in the score
| Benchmark | Result | Source | Why not used |
|---|---|---|---|
| ARC-AGI-1 | 78.0% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 67.0% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 52.5% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 48.5% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 | 5.5% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 78.0% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 67.0% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 52.5% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 48.5% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-1 (ARC Prize verified) | 5.5% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 | 33.1% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 13.6% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 4.9% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 4.0% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 | 0.0% | Epoch AI | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 33.1% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 13.6% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 4.9% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 4.0% | ARC Prize | A stronger configuration result was used |
| ARC-AGI-2 (ARC Prize verified) | 0.0% | ARC Prize | A stronger configuration result was used |
Names matched to this model
inkling-small, inkling-small-peft-262144, inkling-small-peft-262144-sampling-nvfp4