vs
5:5
overall capability wins (0 ties across 10 scored). K2 Horizon 375B A23B leads overall.
- Reasoning2:3
- Coding1:1
- Tool use0:1
- Composite indices0:2
- Specs2:0
Verdict
K2 Horizon 375B A23B edges reasoning & knowledge; Inkling offers ~2.0x larger context (1.0M vs 524K).
- K2 Horizon 375B A23B edges reasoning & knowledge
- Inkling offers ~2.0x larger context (1.0M vs 524K)
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Inkling
- $2.01
- K2 Horizon 375B A23B
- -
Price per 1M tokens
Full comparison
Organization
Thinking Machines
MBZUAI Institute of Foundation Models
Family
Other
Other
License
Apache 2.0
Apache 2.0
Open weights
Yes
Yes
Release
Jul 15, 2026
Sep 3, 2026
Knowledge cutoff
-
-
API / provider
OpenRouter / Tinker (ref.)
No primary API price listed
Modalities
text, image, audio → text
text → text
Specs
Context window
1.0M
524K
Max output
131K
-
Parameters
975B (41B act.)
375B (23B act.)
Pricing
Input $/1M
$1
—
Output $/1M
$4.05
—
Blended $/1M (3∶1)
$1.76
-
Speed
tok/s
55
-
TTFT (s)
0.55
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
87.2
87.3
Humanity's Last Exam
31.9
32.0
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
41.6
18.2
AA-LCR v1.1
77.3
80.0
CritPt
5.4
4.6
MMMU-Pro
73.5
—
IFBench
—
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
77.6
—
SWE-bench Pro
54.3
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
55.1
71.9
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
47.0
42.9
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,410
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,440
—
| Metric | Inkling | K2 Horizon 375B A23B | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Thinking Machines | MBZUAI Institute of Foundation Models | — |
| Family | Other | Other | — |
| License | Apache 2.0 | Apache 2.0 | — |
| Open weights | Yes | Yes | — |
| Release | Jul 15, 2026 | Sep 3, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | OpenRouter / Tinker (ref.) | No primary API price listed | — |
| Modalities | text, image, audio → text | text → text | — |
| Specs | |||
| Context window | 1.0M | 524K | A +524K |
| Max output | 131K | - | — |
| Parameters | 975B (41B act.) | 375B (23B act.) | A +600 |
| Pricing | |||
| Input $/1M | $1 | — | — |
| Output $/1M | $4.05 | — | — |
| Blended $/1M (3∶1) | $1.76 | - | — |
| Speed | |||
| tok/s | 55 | - | — |
| TTFT (s) | 0.55 | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 87.2 | 87.3 | B +0.1 pts |
| Humanity's Last Exam | 31.9 | 32.0 | B +0.1 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 41.6 | 18.2 | A +23.4 pts |
| AA-LCR v1.1 | 77.3 | 80.0 | B +2.7 pts |
| CritPt | 5.4 | 4.6 | A +0.8 pts |
| MMMU-Pro | 73.5 | — | — |
| IFBench | — | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | 77.6 | — | — |
| SWE-bench Pro | 54.3 | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 55.1 | 71.9 | B +16.8 pts |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 47.0 | 42.9 | A +4.1 pts |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,410 | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,440 | — | — |
Inkling
- Context
- 1.0M / 131K out
- Parameters
- 975B (41B act.)
- Price
- $1 / $4.05
- Speed
- 55 tok/s · 0.55s TTFT
- Modalities
- text, image, audio → text
- License
- Apache 2.0
Thinking Machines' first open-weights MoE (975B / 41B active) — natively multimodal with variable thinking effort and 1M context; Apache 2.0 weights on Hugging Face.
K2 Horizon 375B A23B
- Context
- 524K
- Parameters
- 375B (23B act.)
- Price
- —
- Speed
- —
- Modalities
- text → text
- License
- Apache 2.0
Flagship model (375B MoE / 23B active) in the Institute of Foundation Models' K2 Horizon fleet, released fully open — weights, code, training data, and methodology — under Apache 2.0.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.