API access
Subsidized inference in exchange for de-identified reasoning traces. Every call is scrubbed, scored, and tracked.
Create a key
These help us describe the dataset in aggregate (e.g. "traces from legal teams"). They are never exported alongside your company name.
Signup is invite-only during the beta. Your invite gives prepaid inference credit: no card, no charge. The credit is non-refundable, non-transferable and tied to this key.
Available models Choose per request — no signup-time lock-in
| Model | Discount | AA Coding | ||||
|---|---|---|---|---|---|---|
| DeepSeek | ||||||
|
DeepSeek V4 Flash
deepseek-v4-flash-0731
|
$0.060/M$0.006/M | $0.18/M$0.018/M | 90% off | 34.3 | 69.1 | 1M |
|
DeepSeek V4.1 Flash
deepseek-v4p1-flash
|
$0.30/M$0.030/M | $1.20/M$0.12/M | 90% off | 39.5 | — | 1M |
|
DeepSeek V4 Pro
deepseek-v4-pro-0813
|
$1.30/M$0.13/M | $2.60/M$0.26/M | 90% off | 36.0 | 68.8 | 1M |
|
DeepSeek V4 Pro (rolling)
deepseek-v4-pro
|
$1.30/M$0.13/M | $2.60/M$0.26/M | 90% off | 36.0 | 68.8 | 1M |
|
DeepSeek V3
deepseek-v3
|
$0.32/M$0.032/M | $0.89/M$0.089/M | 90% off | 8.5 | 23.0 | 164K |
|
DeepSeek V3.1
deepseek-v3p1
|
$0.25/M$0.025/M | $0.95/M$0.095/M | 90% off | 13.5 | — | 164K |
|
DeepSeek V3.2
deepseek-v3p2
|
$0.26/M$0.026/M | $0.38/M$0.038/M | 90% off | 21.5 | 44.2 | 164K |
|
DeepSeek R1
deepseek-r1
|
$0.50/M$0.050/M | $2.15/M$0.21/M | 90% off | 13.1 | — | 164K |
|
DeepSeek-V4-Flash
deepseek-v4-flash
|
$0.090/M$0.009/M | $0.18/M$0.018/M | 90% off | 34.3 | 69.1 | 1M |
|
DeepSeek-V4-Flash-Vision-Exp
deepseek-v4-flash-vision-exp
|
$0.44/M$0.044/M | $1.32/M$0.13/M | 90% off | 34.8 | 65.0 | 1M |
|
DeepSeek-V4.1-Flash-1017
deepseek-v4.1-flash-1017
|
$0.20/M$0.020/M | $0.60/M$0.060/M | 90% off | 45.1 | — | 1M |
| Qwen | ||||||
|
Qwen3-235B
qwen3-235b
|
$0.090/M$0.009/M | $0.55/M$0.055/M | 90% off | 12.0 | — | 262K |
|
Qwen 3.8 Max
qwen3-max
|
$2.00/M$0.20/M | $6.00/M$0.60/M | 90% off | 40.2 | 71.8 | 262K |
|
Qwen3.8-2.4T-A95B
qwen3p8-2p4t
|
$2.00/M$0.20/M | $6.00/M$0.60/M | 90% off | 39.9 | 71.9 | — |
|
Qwen3.8-27B
qwen3p8-27b
|
$0.40/M$0.040/M | $3.00/M$0.30/M | 90% off | 33.7 | 68.1 | 262K |
|
Qwen3.8 Flash
qwen3p8-flash
|
$0.11/M$0.011/M | $0.38/M$0.038/M | 90% off | — | — | — |
|
Qwen3-14B
qwen3-14b
|
$0.12/M$0.012/M | $0.24/M$0.024/M | 90% off | 6.4 | 13.8 | 41K |
|
Qwen3-30B-A3B
qwen3-30b-a3b
|
$0.12/M$0.012/M | $0.50/M$0.050/M | 90% off | 7.6 | — | 41K |
|
Qwen3-32B
qwen3-32b
|
$0.080/M$0.008/M | $0.28/M$0.028/M | 90% off | 7.2 | 15.3 | 41K |
|
Qwen3-Coder-480B-A35B-Instruct-Turbo
qwen3-coder-480b-a35b-instruct-turbo
|
$0.30/M$0.030/M | $1.00/M$0.10/M | 90% off | 11.9 | — | 262K |
|
Qwen3-Next-80B-A3B-Instruct
qwen3-next-80b-a3b-instruct
|
$0.090/M$0.009/M | $1.10/M$0.11/M | 90% off | 9.6 | — | 262K |
|
Qwen3.5-27B
qwen3.5-27b
|
$0.26/M$0.026/M | $2.60/M$0.26/M | 90% off | 22.9 | — | 262K |
|
Qwen3.5-35B-A3B
qwen3.5-35b-a3b
|
$0.14/M$0.014/M | $1.00/M$0.10/M | 90% off | 19.3 | — | 262K |
|
Qwen3.5-397B-A17B
qwen3.5-397b-a17b
|
$0.45/M$0.045/M | $3.00/M$0.30/M | 90% off | 18.4 | 48.2 | 262K |
|
Qwen3.5-9B
qwen3.5-9b
|
$0.10/M$0.010/M | $0.15/M$0.015/M | 90% off | 11.2 | 28.7 | 262K |
|
Qwen3.6-27B
qwen3.6-27b
|
$0.32/M$0.032/M | $3.20/M$0.32/M | 90% off | 21.4 | 53.7 | 262K |
|
Qwen3.6-35B-A3B
qwen3.6-35b-a3b
|
$0.10/M$0.010/M | $0.95/M$0.095/M | 90% off | 18.2 | 41.9 | 262K |
| Moonshot / Kimi | ||||||
Kimi K2.6
kimi-k2p6
|
$0.75/M$0.075/M | $3.50/M$0.35/M | 90% off | 27.0 | 61.8 | 262K |
Kimi K3
kimi-k3
|
$2.85/M$0.28/M | $14.25/M$1.43/M | 90% off | 43.6 | 76.2 | 1M |
| Zhipu / GLM | ||||||
GLM-5.2
glm-5p2
|
$0.75/M$0.075/M | $2.40/M$0.24/M | 90% off | 33.7 | 68.8 | 1M |
GLM-5.3
glm-5p3
|
$1.40/M$0.14/M | $4.40/M$0.44/M | 90% off | 44.8 | 74.8 | 1M |
GLM-5.3 Flash
glm-5p3-flash
|
$0.15/M$0.015/M | $0.50/M$0.050/M | 90% off | 41.8 | 71.5 | 1M |
GLM-4.6
glm-4p6
|
$0.50/M$0.050/M | $2.00/M$0.20/M | 90% off | 18.5 | 45.8 | 203K |
GLM-4.7
glm-4p7
|
$0.40/M$0.040/M | $1.75/M$0.17/M | 90% off | 22.2 | 45.3 | 203K |
| Thinking Machines Lab | ||||||
Inkling (Flagship)
inkling
|
$1.00/M$0.10/M | $4.05/M$0.41/M | 90% off | 25.0 | 52.1 | 1M |
| OpenAI (open-weight) | ||||||
GPT-OSS 120B
gpt-oss-120b
|
$0.037/M$0.004/M | $0.17/M$0.017/M | 90% off | 11.6 | 30.4 | 131K |
GPT-OSS 20B
gpt-oss-20b
|
$0.030/M$0.003/M | $0.14/M$0.014/M | 90% off | 9.0 | 20.7 | 131K |
| Mistral | ||||||
|
Mistral Small 3.2 (24B)
mistral-small-3p2-24b
|
$0.075/M$0.007/M | $0.20/M$0.020/M | 90% off | 7.0 | 12.5 | 128K |
| Meta | ||||||
Muse Spark Open
muse-spark-open
Built with Muse Spark Open
|
$1.25/M$0.12/M | $4.25/M$0.42/M | 90% off | 38.9 | — | 1M |
| Xiaomi / MiMo | ||||||
|
MiMo-V2.6-Pro
mimo-v2p6-pro
|
$0.43/M$0.043/M | $0.87/M$0.087/M | 90% off | 46.3 | — | — |
|
MiMo-V2.6-Flash
mimo-v2p6-flash
|
$0.14/M$0.014/M | $0.28/M$0.028/M | 90% off | 37.9 | — | — |
| MiniMax | ||||||
|
MiniMax M3
minimax-m3
|
$0.28/M$0.028/M | $1.10/M$0.11/M | 90% off | 29.2 | 58.6 | 524K |
| NVIDIA / Nemotron | ||||||
|
Nemotron 3 Ultra
nemotron-3-ultra
|
$0.50/M$0.050/M | $2.20/M$0.22/M | 90% off | 22.9 | 49.3 | 262K |
AA Index and AA Coding: independent evaluations by Artificial Analysis; hover a score for the variant it ran.
Open-weight models carry their makers' licenses; some require attribution or a separate agreement at scale. See Model licenses.
Elasticity holds the underlying provider credentials — you never see or manage a provider key.
Look up usage
Test the scrubbing pipeline
Paste text or upload a single file — the real 3-pass pipeline: Pass 1 (pattern match), Pass 2 (entity recognition), and a live model audit that catches what the first two miss. Nothing you enter is stored.