Plan
Your subscription, GPU allocation, and upgrade options.
Tier 2 — Managed Inference
Charter
Rate locked forever
Professional-Grade GPU · 20 GB VRAM · Houston, TX · 12-month term · ROS-01-I
$999/mo
Renews Jun 14, 2026 · $0 token fees
What's Included
Base · Hardware
✓ Dedicated GPU — yours alone, no neighbors
✓ IPMI / out-of-band access
✓ OpenAI-compatible endpoint
✓ Your data never leaves Houston TX
✓ Swap API key, drop in replacement
✓ 20 GB VRAM, 1 TB data volume
Tier 1 · Managed Infrastructure
✓ OS lifecycle (Ubuntu 24.04 LTS)
✓ NVIDIA driver + CUDA, version-pinned
✓ 24/7 NOC — hardware alerts & response
✓ PSU, GPU thermal, ECC error monitoring
✓ Daily incremental + weekly full backup
✓ 4-hr on-site break/fix (business hours)
Tier 2 · Managed Inference Platform — Your current tier
✓ vLLM deployed, configured & tuned
✓ Model deployment automation (HF ID or S3)
✓ AWQ/GPTQ/FP8 quantization pipeline
✓ API-key auth + per-key token accounting
✓ Traffic balancing — no agent bottlenecks
✓ P95 TTFT SLO · model rollback on error
Upgrade Options
Regulated Inference
Tier 2 + full compliance stack. HIPAA BAA, SOC 2, audit logs. Required for healthcare, finance, and legal workloads.
✓ HIPAA BAA executed & active
✓ SOC 2 Type II report inheritance
✓ 7-year tamper-evident audit log
✓ Dedicated VLAN + US-persons-only support
✓ Annual third-party pen test report
+$999/mo
Upgrade to Tier 3
Full AI infrastructure partnership. Named TAM, fine-tuning, RAG ops, and quarterly model reviews.
✓ Quarterly model selection review
✓ LoRA/QLoRA fine-tuning support
✓ RAG stack deployed & operated (pgvector / Qdrant)
✓ Drift detection + eval harness
✓ Named TAM + monthly business review
+$1,499/mo
Add GPU
A second dedicated GPU at the same Charter rate. Run 30B+ models or isolate prod and dev inference.
✓ Same professional-grade GPU hardware
✓ Assigned a permanent physical ID — ROS-01-II
✓ Charter rate locked — $999/mo
✓ Runs independently or load-balanced
✓ Can run 30B+ AWQ with 2× 20 GB VRAM
$999/mo per GPU
Billing History
| Period | Plan | Amount | Status |
|---|---|---|---|
| May 2026 | Tier 2 — Managed Inference | $999 | Due Jun 1 |
| Apr 2026 | Tier 2 — Managed Inference | $999 | Paid |
| Mar 2026 | Tier 2 — Managed Inference | $999 | Paid |
Overview
May 2026 — your savings at a glance.
This Month — May 2026
Your plan
Tier 2 · $999/mo
Anthropic equivalent
$3,360
You saved
$2,361 70%
560M tokens · 420M input × $3.00/M + 140M output × $15.00/M = $3,360 Anthropic equivalent
Your flat monthly plan: $999 · Savings: $2,361 (70%)
Energy arbitrage is priced into your plan rate — no separate power bill. Also vs. GPT-4o: $2,155 saved
Your Plan
Tier 2 — Managed Inference
Professional-Grade GPU · 20 GB VRAM · Houston TX · ROS-01-I
Charter
Rate locked
$0 token fees
Model slot capacity (20 GB VRAM)
7B class modelsup to 2 concurrent
14B AWQ models1 at a time
30B+ modelsrequires upgrade
Cost per Million Tokens
$1.61/M
Active Agents
14
↑2 from last week
API Calls This Month
18,432
↑18% WoW
GPU Allocation Used
23% of 1 GPU
77% free
Rate Limit Incidents
0
GPU Status
Operational
24-Hour Token Activity
Refreshes every 5 min
0:003:006:009:0012:0015:0018:0021:00
Usage
Token consumption, spend, and agent activity.
Effective Rate This Month
$1.61/M tokens
vs. $6.00/M on Anthropic (blended)
Plan Fee This Month
$999
560M tokens · 18,432 API calls · $0 token fees
Anthropic Equivalent
$3,360
$2,361 saved this month
Daily Token Usage — May 2026
May 1May 7May 14May 21May 28
By Model
| Model | Tokens | Cost | Rate | vs. Anthropic | Savings |
|---|---|---|---|---|---|
| Qwen2.5-14B-Instruct-AWQ | 448M | $734 | $1.64/M | $3,136 | $2,402 |
| Qwen2.5-7B-Instruct-AWQ | 112M | $169 | $1.51/M | $784 | $615 |
By Agent
| Agent ID | Tokens | Calls | Cost | Rate |
|---|---|---|---|---|
| agent_01JX4K2P… | 210M | 7,200 | $340 | $1.62/M |
| agent_01JX3M8R… | 168M | 5,541 | $272 | $1.62/M |
| agent_01JX1Q9T… | 112M | 3,686 | $181 | $1.62/M |
| agent_01JWZN5V… | 70M | 2,005 | $110 | $1.57/M |
Rolling 30-day window. Month-to-date totals on "This Month" align with your billing cycle.
By Model — Last 30 Days
| Model | Tokens | Cost | Rate | Savings vs. Anthropic |
|---|---|---|---|---|
| Qwen2.5-14B-Instruct-AWQ | 512M | $838 | $1.64/M | $2,746 |
Billing History
| Month | Tokens | Effective Rate | Plan Fee | Anthropic Equiv. | Savings |
|---|---|---|---|---|---|
| May 2026 | 560M | $1.61/M | $999 | $3,360 | $2,361 |
| Apr 2026 | 498M | $1.64/M | $999 | $2,934 | $1,935 |
| Mar 2026 | 445M | $1.64/M | $999 | $2,620 | $1,621 |
Plan fee is flat $999/mo — $0 per-token fees on all calls regardless of volume.
Models
Models running on your GPU and those available to deploy.
Running Now
Qwen/Qwen2.5-14B-Instruct-AWQ
LIVE
14B params · AWQ 4-bit · General purpose
VRAM
18.4 / 20 GB
Model Library — Available to Deploy
Qwen/Qwen2.5-7B-Instruct-AWQ
mistralai/Mistral-7B-Instruct-v0.3
meta-llama/Llama-3.1-8B-Instruct
microsoft/Phi-4
google/gemma-2-9b-it
deepseek-ai/DeepSeek-R1-Distill-Qwen-7B
Qwen/Qwen2.5-14B-Instruct-AWQ
Don't see your model?
Request a custom model
Any model on HuggingFace. We'll evaluate VRAM fit and deploy within 4 business hours if it's compatible with your GPU allocation.
API Keys
Manage keys for authenticating to the BotINFRA API.
Your Keys
| Name | Key | Created | Last Used | Calls / Month | |
|---|---|---|---|---|---|
| Production | bi_live_•••••••• | Jan 14, 2025 | May 8, 2026 | 14,210 | |
| Staging | bi_live_•••••••• | Mar 3, 2026 | May 7, 2026 | 4,222 |
Create New Key
Status
Is BotINFRA working for you right now?
Uptime This Month
99.98%
7 min downtime
Rate Limit Incidents
0
No throttled requests
Scheduled Maintenance
None
No planned outages
Response Time SLA
| Model | P95 Commitment | This Month (avg) | Status |
|---|---|---|---|
| Qwen2.5-14B (current) | 500 ms first token | Monitoring pending | Coming soon |
| 7B class models | 300 ms first token | Monitoring pending | Coming soon |
Latency instrumentation is in development. SLA credits apply retroactively once active.
Incident History — Last 90 Days
| Date | Description | Duration | Status |
|---|---|---|---|
| May 1, 2026 | Scheduled GPU driver update | 7 min | Resolved |
| No other incidents in the past 90 days. | |||
Compliance
Your data, your control. Proof for legal and IT.
Demo:
All 5 controls active · $999/mo Regulated Inference
HIPAA BAA
Active
Business Associate Agreement signed Jan 14, 2025. Your agents can process protected health information (PHI).
SOC 2 Type II
Compliant
Last audited March 2026. Report available on request. Covers security, availability, and confidentiality.
Data Residency
Confirmed
All inference runs on hardware in Houston, TX, USA. Your prompts and outputs never leave this facility. Included on all plans.
Zero Data Retention
Active
Inference requests and responses are not stored or logged for model training. No prompts retained after the API call completes. Included on all plans.
Audit Log Retention
7-Year Retention
Every inference call logged with timestamp, agent ID, model, and token count. Tamper-evident storage. Downloadable for legal review.
Encryption & BYOK
Active
Encryption at rest verified. Customer-managed keys (BYOK) available. US-persons-only support staffing, background-checked.
Annual Pen Test
Last: Feb 2026
Annual third-party penetration test. Report available to your security team on request. Next test: Feb 2027.
Audit Log
| Timestamp | Agent ID | Model | Input Tokens | Output Tokens |
|---|---|---|---|---|
| 2026-05-08 14:31:02Z | agent_01JX4K2P | Qwen2.5-14B-AWQ | 4,218 | 832 |
| 2026-05-08 14:29:44Z | agent_01JX3M8R | Qwen2.5-14B-AWQ | 2,104 | 641 |
| 2026-05-08 14:28:11Z | agent_01JX4K2P | Qwen2.5-14B-AWQ | 6,540 | 1,204 |
| 2026-05-08 14:25:58Z | agent_01JX1Q9T | Qwen2.5-14B-AWQ | 1,882 | 490 |
| 2026-05-08 14:22:30Z | agent_01JWZN5V | Qwen2.5-14B-AWQ | 3,310 | 720 |
Showing 5 most recent. Export for full log — retained 7 years.
Agents
Live health board — status, errors, idle detection, and spend attribution across your fleet.
Total Agents
14
↑2 from last week
Running Now
12
Idle (>2h)
2
Consider pausing
Avg Error Rate
0.2%
Well within SLO
Calls Today
2,841
↑22% vs yesterday
Cost Today
$34.21
On pace: $1,027/mo
⚠️
2 agents haven't made a call in over 2 hours.
Doc Summarizer and Report Builder may be stuck or complete. Review and pause if unneeded.
Agent Fleet — Live
Refreshes every 30s
| Agent | Role | Status | Last Call | Calls Today | Error Rate | Cost Today | WoW Trend |
|---|---|---|---|---|---|---|---|
| agent_01JX4K2P | Outreach Bot | Running | 2 min ago | 2,841 | 0.0% | $14.20 | ↑12% |
| agent_01JX3M8R | Research Bot | Running | 5 min ago | 1,892 | 0.4% | $9.18 | ↑8% |
| agent_01JX1Q9T | Data Extractor | Running | 8 min ago | 1,210 | 0.0% | $5.92 | ↑22% |
| agent_01JW8K4R | Support Triage | Running | 1 min ago | 1,544 | 0.8% ↑ | $7.21 | ↑5% |
| agent_01JWZN5V | Lead Scorer | Running | 12 min ago | 764 | 0.0% | $3.44 | ↓3% |
| agent_01JW2P4M | Intake Classifier | Running | 3 min ago | 488 | 0.0% | $2.14 | ↑34% |
| agent_01JV9K8T | Proposal Writer | Running | 18 min ago | 312 | 0.0% | $1.48 | ↑17% |
| agent_01JV5M2T | Doc Summarizer | Idle 3h | 3 hr ago | 48 | 0.0% | $0.21 | ↓82% |
| agent_01JV3N8P | Report Builder | Idle 5h | 5 hr ago | 12 | 0.0% | $0.05 | ↓91% |
Showing 9 of 14 active agents. All agents share the flat $999/mo plan — cost attribution is for internal tracking only.
Spend Attribution — This Month
Cost by Agent — May 2026
| Agent | Role | Tokens | Calls | % of Fleet | Attributed Cost |
|---|---|---|---|---|---|
| agent_01JX4K2P | Outreach Bot | 210M | 7,200 | 37.5% | $374 |
| agent_01JX3M8R | Research Bot | 168M | 5,541 | 30.0% | $300 |
| agent_01JX1Q9T | Data Extractor | 112M | 3,686 | 20.0% | $200 |
| agent_01JWZN5V | Lead Scorer | 70M | 2,005 | 12.5% | $125 |
Attributed cost = agent's % of total tokens × $999 plan fee. Useful for internal chargeback or ROI analysis.
Impact
The environmental and research contributions your compute is making.
CO₂e Avoided This Month
19.7
metric tons
vs. drawing equivalent power from ERCOT grid
51,100 kWh × 0.386 kg CO₂e/kWh ÷ 1,000 = 19.72 MT
Source: EPA eGRID 2023, Texas ERCT subregion
Equivalent to removing ~4.3 cars from the road for one year
Lifetime avoided: 118 MT
Health Research Funded This Month
$1,050
contributed
to the American Open-Source Health Research Endowment
35% of your subscription above BotINFRA's operating threshold
This month: $1,050
Lifetime total: $8,925
Funds subsidized compute for American biomedical AI researchers — open-source, publicly accessible results.
Monthly CO₂e Avoided (MT)
JanFebMarAprMay
Research Funding — Lifetime Allocation
| Research Area | Institution | Funded |
|---|---|---|
| Cancer genomics compute | MD Anderson Cancer Center | $2,850 |
| Drug discovery LLMs | Baylor College of Medicine | $2,310 |
| Radiology AI training | UTHealth Houston | $1,890 |
| Rare disease classification | Texas Children's Hospital | $1,050 |
| General endowment reserve | AOHRE | $825 |
Allocations are made quarterly by the AOHRE board. All funded research is published open-access.