vs
10:0
overall capability wins (0 ties across 10 scored). Inkling-Small leads overall.
- Reasoning5:0
- Coding3:0
- Tool use1:0
- Composite indices2:0
- Pricing0:3
- Specs1:0
Verdict
Inkling-Small leads coding (widest gap: +240.0 on WebDev Elo); Inkling-Small offers ~8.2x larger context (1.0M vs 128K); Inkling-Small ships open weights (Apache 2.0).
- Inkling-Small leads coding (widest gap: +240.0 on WebDev Elo)
- Inkling-Small offers ~8.2x larger context (1.0M vs 128K)
- Inkling-Small ships open weights (Apache 2.0)
- Mercury 2 is ~2.1x cheaper on a blended token basis than Inkling-Small
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Inkling-Small
- $0.94
- Mercury 2
- $0.44
Mercury 2 is about 2.1x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Thinking Machines
Inception Labs
Family
Other
Other
License
Apache 2.0
Proprietary
Open weights
Yes
No
Release
Jul 30, 2026
Sep 1, 2026
Knowledge cutoff
-
-
API / provider
Tinker (ref.)
Inception Labs
Modalities
text, image, audio → text
text → text
Specs
Context window
1.0M
128K
Max output
131K
-
Parameters
276B (12B act.)
—
Pricing
Input $/1M
$0.58
$0.25
Output $/1M
$1.44
$0.75
Blended $/1M (3∶1)
$0.79
$0.38
Speed
tok/s
85
-
TTFT (s)
0.35
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
89.5
77.0
Humanity's Last Exam
33.3
17.1
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
33.2
21.2
AA-LCR v1.1
75.7
43.7
CritPt
8.3
0.8
MMMU-Pro
74.0
—
IFBench
—
69.8
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
80.2
—
SWE-bench Pro
55.9
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
55.1
27.3
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
49.7
37.7
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,407
1,167
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,405
—
| Metric | Inkling-Small | Mercury 2 | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Thinking Machines | Inception Labs | — |
| Family | Other | Other | — |
| License | Apache 2.0 | Proprietary | — |
| Open weights | Yes | No | — |
| Release | Jul 30, 2026 | Sep 1, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Tinker (ref.) | Inception Labs | — |
| Modalities | text, image, audio → text | text → text | — |
| Specs | |||
| Context window | 1.0M | 128K | A +921K |
| Max output | 131K | - | — |
| Parameters | 276B (12B act.) | — | — |
| Pricing | |||
| Input $/1M | $0.58 | $0.25 | B +$0.33 |
| Output $/1M | $1.44 | $0.75 | B +$0.69 |
| Blended $/1M (3∶1) | $0.79 | $0.38 | B +$0.42 |
| Speed | |||
| tok/s | 85 | - | — |
| TTFT (s) | 0.35 | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 89.5 | 77.0 | A +12.5 pts |
| Humanity's Last Exam | 33.3 | 17.1 | A +16.2 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 33.2 | 21.2 | A +12.0 pts |
| AA-LCR v1.1 | 75.7 | 43.7 | A +32.0 pts |
| CritPt | 8.3 | 0.8 | A +7.5 pts |
| MMMU-Pro | 74.0 | — | — |
| IFBench | — | 69.8 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | 80.2 | — | — |
| SWE-bench Pro | 55.9 | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 55.1 | 27.3 | A +27.8 pts |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 49.7 | 37.7 | A +12.0 pts |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,407 | 1,167 | A +240 Elo |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,405 | — | — |
Inkling-Small
- Context
- 1.0M / 131K out
- Parameters
- 276B (12B act.)
- Price
- $0.58 / $1.44
- Speed
- 85 tok/s · 0.35s TTFT
- Modalities
- text, image, audio → text
- License
- Apache 2.0
Efficient Inkling sibling (276B / 12B active MoE) — matches or beats Inkling on many benchmarks at ~¼ the size; Apache 2.0 weights with Tinker fine-tuning access.
Mercury 2
- Context
- 128K
- Parameters
- —
- Price
- $0.25 / $0.75
- Speed
- —
- Modalities
- text → text
- License
- Proprietary
Inception Labs' diffusion-based (non-autoregressive) reasoning LLM, API-only with a 128K context window and high-throughput inference on Blackwell GPUs.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.