Inkling 2 Mini and Phi-4-Reasoning-Vision-15B have no benchmarks in common in this dataset, so their scores cannot be compared directly. The specification rows below still apply.
Price basis: Standard rates vs listed rates. The price section below shows the applicable periods.
| Fact | Inkling 2 Mini | Phi-4-Reasoning-Vision-15B |
|---|---|---|
| Context window | 1M tokens | 16K tokens |
| Input / 1M | $0.150 | — |
| Output / 1M | $0.500 | — |
| Blended (4:1) | $0.123 | — |
| Status | Available | Available |
| Released | November 20, 2026 | March 4, 2026 |
Blended is 4:1 input to output with 90% cache hits at listed rates — set your own mix.
Published scores, with matching modes where available. Gaps show Inkling 2 Mini relative to Phi-4-Reasoning-Vision-15B. Reading the results
A dash means no result is available.
| Benchmark | Inkling 2 Mini | Phi-4-Reasoning-Vision-15B |
|---|---|---|
| HLE-Diamond | 15.2 | — |
| DeepSWE | 48.2 | — |
| FrontierCode 2.0 | 15.1 | — |
| Vibe Code Bench v1.1 | 77.9 | — |
| AutomationBench-AA | 41.8 | — |
| MCP-Atlas | 82.7 | — |
| ScreenSpot-Pro (no tools) | 76.5 | — |
| ScreenSpot-v2 | — | 88.3 |
| AA Intelligence Index v5.0.1 | 20.1 | — |
| Vals Index v2.2 | 46.5 | — |
| LongBench v2 | 66.9 | — |
| OpenAI MRCR v2 (8-needle, 256K–512K) | 46.4 | — |
| AI2D | — | 84.7 |
| Chartography | 72.5 | — |
| ChartQA | — | 82.9 |
| GDP.pdf | 16.8 | — |
| HallusionBench | — | 63.9 |
| MathVerse (testmini) | — | 53.1 |
| MathVision (testmini) | — | 36.2 |
| MathVista (testmini) | — | 74.1 |
| MMMU | — | 55.0 |
| MMMU-Pro | 74.9 | — |
| MMStar | — | 63.9 |
| OCRBench | — | 75.6 |
| OmniDocBench | 92.8 | — |
| PerceptionBench | 57.5 | — |
| ZeroBench Main | 34.0 | — |
| AA-AnalystAgent | 46.3 | — |
| AA-Briefcase | 1263 | — |
| GDPval-AA v2 | 1319 | — |
| JobBench | 49.6 | — |
| ARC-AGI-1 | 84.5 | — |
| ARC-AGI-2 | 49.3 | — |
| BrowseComp | 79.0 | — |
| DeepSearchQA | 85.3 | — |
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix · 5% steps
Blended price = effective input rate at 90% cache hits × 80% + output rate × 20%
A blended cost needs a published input and output price on both sides, and one of these two does not have them yet.
Inkling 2 Mini: Standard pricing. Tinker 64K/256K training (1M sampling); fine-tuning billed separately ($0.75/1M for 64K, $1.50/1M for 256K).
Phi-4-Reasoning-Vision-15B: Listed text rates; pricing tier not confirmed.
| Price / 1M tokens | Inkling 2 Mini | Phi-4-Reasoning-Vision-15B | Ratio |
|---|---|---|---|
| Input | $0.15 | — | — |
| Output | $0.50 | — | — |
| Cached input | $0.015 | — | — |
| Cache write | — | — | — |
| Blended | $0.123 | — | — |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
| Attribute | Inkling 2 Mini | Phi-4-Reasoning-Vision-15B |
|---|---|---|
| Availability | Available | Available |
| Provider | Thinking Machines Lab | Microsoft |
| Model type | Reasoning model | Reasoning model |
| First release | November 20, 2026 | March 4, 2026 |
| Context window | 1M tokens | 16K tokens |
| Total parameters | 26.4B | 15B |
| Active parameters | 5.4B | — |
| Architecture | Mixture of Experts (MoE) | Mixture of Experts (MoE) |
| Documented capabilities | Reasoning, Coding, Tool use, Long context | Reasoning, OCR, Computer use |
| Input modalities | Text, Image | Text, Image |
| Output modalities | Text | Text |
| Weights | Open weights | Open weights |
| Licence | Apache 2.0 | MIT License |
| Commercial use | Commercial use allowed | Commercial use allowed |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used Inkling 2 Mini or Phi-4-Reasoning-Vision-15B? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.