vs
6:0
overall capability wins (0 ties across 6 scored). Agnes-3.0-Flash Preview leads overall.
- Reasoning4:0
- Coding1:0
- Specs1:0
Verdict
Agnes-3.0-Flash Preview leads coding (widest gap: +0.4 on SciCode); Agnes-3.0-Flash Preview offers ~2.0x larger context (262K vs 128K); Agnes-3.0-Flash Preview ships open weights (Apache 2.0).
- Agnes-3.0-Flash Preview leads coding (widest gap: +0.4 on SciCode)
- Agnes-3.0-Flash Preview offers ~2.0x larger context (262K vs 128K)
- Agnes-3.0-Flash Preview ships open weights (Apache 2.0)
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Agnes-3.0-Flash Preview
- -
- Mercury 2
- $0.44
Price per 1M tokens
Full comparison
Organization
Agnes AI
Inception Labs
Family
Other
Other
License
Apache 2.0
Proprietary
Open weights
Yes
No
Release
Sep 13, 2026
Sep 1, 2026
Knowledge cutoff
-
-
API / provider
No primary API price listed
Inception Labs
Modalities
text, image, video → text
text → text
Specs
Context window
262K
128K
Max output
-
-
Parameters
33B
—
Pricing
Input $/1M
—
$0.25
Output $/1M
—
$0.75
Blended $/1M (3∶1)
-
$0.38
Speed
tok/s
-
-
TTFT (s)
-
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
85.0
77.0
Humanity's Last Exam
—
17.1
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
23.0
21.2
AA-LCR v1.1
68.3
43.7
CritPt
—
0.8
MMMU-Pro
—
—
IFBench
74.2
69.8
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
—
27.3
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
38.1
37.7
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
1,167
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
—
| Metric | Agnes-3.0-Flash Preview | Mercury 2 | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Agnes AI | Inception Labs | — |
| Family | Other | Other | — |
| License | Apache 2.0 | Proprietary | — |
| Open weights | Yes | No | — |
| Release | Sep 13, 2026 | Sep 1, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | No primary API price listed | Inception Labs | — |
| Modalities | text, image, video → text | text → text | — |
| Specs | |||
| Context window | 262K | 128K | A +134K |
| Max output | - | - | — |
| Parameters | 33B | — | — |
| Pricing | |||
| Input $/1M | — | $0.25 | — |
| Output $/1M | — | $0.75 | — |
| Blended $/1M (3∶1) | - | $0.38 | — |
| Speed | |||
| tok/s | - | - | — |
| TTFT (s) | - | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 85.0 | 77.0 | A +8.0 pts |
| Humanity's Last Exam | — | 17.1 | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 23.0 | 21.2 | A +1.8 pts |
| AA-LCR v1.1 | 68.3 | 43.7 | A +24.6 pts |
| CritPt | — | 0.8 | — |
| MMMU-Pro | — | — | — |
| IFBench | 74.2 | 69.8 | A +4.4 pts |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | — | 27.3 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 38.1 | 37.7 | A +0.4 pts |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | 1,167 | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | — | — |
Agnes-3.0-Flash Preview
- Context
- 262K
- Parameters
- 33B
- Price
- —
- Speed
- —
- Modalities
- text, image, video → text
- License
- Apache 2.0
Agnes AI's (Sapiens AI, Singapore) first open-weight release: a 33B-parameter hybrid-attention model (54 gated delta-rule recurrent layers alternating with 18 global-attention layers, 3:1) under Apache 2.0, with a 262K-token context window and text/image/video input. This open-weight Preview checkpoint is distinct from Agnes AI's production "Agnes 3.0 Flash" API model, which runs a different checkpoint and is not open-weight; the two should not be conflated.
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.