# TestedActually newsletter, issue 003: Twelve GTM tools benchmarked, and Mixmax drops its cheapest tier

> Our biggest ship yet: the GTM expansion, two new clusters covering the CRM and marketing-automation stack with crawler benchmarks and head-to-heads for the highest-demand pairs. Alongside it, pricing guides for opencode and Fathom and for the local-LLM desktop apps, a Bland voice-AI ladder we verified to the enterprise tier, and six dated market events from the news scan, including Mixmax's entry price moving from $29 to $49 a user.

Human page: https://testedactually.com/newsletter/2026-10-06-gtm-expansion-and-market-events/
Mirror: https://testedactually.com/newsletter/2026-10-06-gtm-expansion-and-market-events.md
Published: 2026-10-06

## In this issue

- **[Ollama pricing, free local models plus a $20 cloud tier](https://testedactually.com/local-llm/ollama/)** (Comparison) - The local-LLM staple keeps local use free forever; the cloud plans run Pro at $20 a month ($16.67 billed annually, $60 of monthly usage credits), Max at $100 ($300 of credits), and Team at $500 ($1,000 of shared credits) [1]. Off-peak rates and per-credit details are in the guide. Mirror: https://testedactually.com/local-llm/ollama.md
- **[Bland - pricing, features, and fit](https://testedactually.com/ai-voice-agents/bland/)** (Comparison) - Bland repositioned upmarket this year, so every claim is re-sourced to its live site and the full ladder is public: Start is free with call time at $0.14 a minute, Build drops that to $0.12 behind a $299 monthly platform fee, Scale runs $0.11 at $499 a month, and Enterprise is contracted to your volume [2]. Mirror: https://testedactually.com/ai-voice-agents/bland.md
- **[Best CRM platforms for startups (HubSpot, Pipedrive, Attio, Close, folk, Zoho)](https://testedactually.com/crm/)** (Comparison) - Six CRM platforms compared on visual pipelines, native email sync, bundled marketing, and free tiers. HubSpot is the platform default with a real free tier (2 users, 1,000 contacts); Pipedrive owns the single-pipeline job from US$14 per seat; Attio bends its data model to your business. HubSpot Professional's $3,000 mandatory onboarding fee is on its own pricing page. Mirror: https://testedactually.com/crm.md
- **[Best marketing automation platforms (Klaviyo, Omnisend, Customer.io, ActiveCampaign, Mailchimp, Braze)](https://testedactually.com/marketing-automation/)** (Comparison) - Six marketing automation and lifecycle platforms split by where their triggers live: storefront (Klaviyo, Omnisend, Mailchimp), product events (Customer.io), builder-first (ActiveCampaign), or enterprise cross-channel (Braze). Free tiers: three are real, one of those has no automation flows, and Braze publishes no prices at all. Mirror: https://testedactually.com/marketing-automation.md
- **[AI crawler access test: can answer engines read these vendors?](https://testedactually.com/crm/benchmarks/crm-ai-crawler-access/)** (Original data) - Six CRM vendors probed three times each on 2026-09-30. Attio and folk block all seven AI crawlers while selling themselves as the CRM for agentic revenue; HubSpot, Pipedrive, Close, and Zoho are fully open. Five of six ship an llms.txt. Mirror: https://testedactually.com/crm/benchmarks/crm-ai-crawler-access.md
- **[AI crawler access test: can answer engines read these vendors?](https://testedactually.com/marketing-automation/benchmarks/marketing-automation-ai-crawler-access/)** (Original data) - The whole marketing-automation category is open to AI crawlers: zero robots blocks across all six vendors, llms.txt on five of six. If LLMs under-recommend a tool in this category, it is not a robots problem. Mirror: https://testedactually.com/marketing-automation/benchmarks/marketing-automation-ai-crawler-access.md
- **[Pipedrive vs HubSpot - pipeline or platform](https://testedactually.com/crm/pipedrive-vs-hubspot/)** (Head to head) - Pipeline or platform. Pipedrive's bill grows with seats (US$14 to start); HubSpot's grows with seats, contacts, hubs, and onboarding fees, and its free email sync is the zero-cost logging layer Pipedrive only matches at US$39 per seat. Mirror: https://testedactually.com/crm/pipedrive-vs-hubspot.md
- **[Klaviyo vs Mailchimp - ecommerce engine or email app](https://testedactually.com/marketing-automation/klaviyo-vs-mailchimp/)** (Head to head) - The entry bands tie at $20/mo, then diverge on what is inside: Klaviyo's free plan runs flows and includes $5 of mobile messages, Mailchimp's free plan includes no automation flows at all. Mailchimp's own homepage advertises switching savings of up to 42 percent, which tells you who it is fighting. Mirror: https://testedactually.com/marketing-automation/klaviyo-vs-mailchimp.md
- **[Jan vs LM Studio - the open-source app or the model workbench](https://testedactually.com/local-llm/jan-vs-lm-studio/)** (Head to head) - The two open-source desktop apps for running local models, compared on licence, interface, model management, and API serving. Jan if you want a free MIT-licensed app that stays out of the way; LM Studio if you want the deeper workbench and its bundled Bionic agent, and will confirm the workplace licence term first. Mirror: https://testedactually.com/local-llm/jan-vs-lm-studio.md
- **[opencode pricing, free software with a pay-per-request model gateway](https://testedactually.com/ai-coding-assistants/opencode/)** (Comparison) - opencode costs nothing to install or run: the agent is free and open source and connects to 75+ model providers you bring yourself. The only money path is Zen, an optional model gateway billing per request at provider rates with the markup capped at processing fees. Mirror: https://testedactually.com/ai-coding-assistants/opencode.md
- **[Fathom pricing, a real free plan and paid tiers from $15 a user](https://testedactually.com/agentic-productivity/fathom/)** (Comparison) - The meeting notetaker with a free plan that works rather than teases: unlimited recordings, transcriptions, and AI summaries for individuals. Paid tiers add depth from $15 a user (Team, annual) to $20 a month Premium, and the guide marks where the free plan genuinely ends. Mirror: https://testedactually.com/agentic-productivity/fathom.md
- **[Mixmax retires its SMB tier, leaving Copilot and Suite only](https://testedactually.com/news/mixmax-drops-smb-tier/)** (News) - Mixmax's pricing page now lists Copilot at $49 and Suite at $89 per user per month, and no longer offers the $29 SMB tier that appeared in our 2026-09-24 check. The cheapest listed path into paid Mixmax is now $49 a user. Mirror: https://testedactually.com/news/mixmax-drops-smb-tier.md
- **[Aleph Alpha ships Kolibri: Apache 2.0 open-weight MoE with 1M-token context](https://testedactually.com/news/aleph-alpha-kolibri-open-weight/)** (News) - Aleph Alpha released Kolibri, an English-German Mixture-of-Experts model with 78.1B total parameters (3.46B active), full weights on Hugging Face under Apache 2.0, and 262K native context extending to 1M via a serve-time flag. A permissive, self-hostable option for EU-data-residency stacks. Mirror: https://testedactually.com/news/aleph-alpha-kolibri-open-weight.md
- **[DeepKeep ships AI Lens: a security control plane for coding agents](https://testedactually.com/news/deepkeep-ai-lens-coding-agents/)** (News) - DeepKeep announced AI Lens for Developers: policy enforcement, audit visibility, and runtime security over coding agents, supporting Cursor and Claude Code today. The coding-agent security lane is forming fast, two weeks after StackHawk's Wingman landed at $10 per user per month. Mirror: https://testedactually.com/news/deepkeep-ai-lens-coding-agents.md
- **[Underdog launches a fully local personal AI agent, no cloud](https://testedactually.com/news/underdog-local-personal-agent/)** (News) - Underdog announced a fully on-device personal AI agent: it reads email, takes meeting notes, handles dictation, manages the calendar, and drives the browser, asking before it pays, with nothing uploaded. Invite-only beta, no pricing disclosed. Mirror: https://testedactually.com/news/underdog-local-personal-agent.md
- **[Deepslate raises €7.7M seed for European speech-to-speech models](https://testedactually.com/news/deepslate-eu-speech-to-speech-seed/)** (News) - Berlin voice-AI lab Deepslate raised a EUR 7.7 million seed led by 42CAP, claiming 440 ms speech-to-speech latency. More model choice and a potential EU-hosted vendor for voice-agent platform buyers at a time when data residency is a real procurement question. Mirror: https://testedactually.com/news/deepslate-eu-speech-to-speech-seed.md

New issues land here first. Subscribe at the site footer; every comparison cited above carries dated, sourced claims.
## References

- [1] Ollama cloud pricing - https://ollama.com/cloud (vendor pricing, checked 2026-10-06)
- [2] Bland pricing page - https://www.bland.ai/pricing (vendor pricing, checked 2026-10-06)
- [3] Mixmax pricing - https://mixmax.com/pricing (live page fetched 2026-10-04: Copilot and Suite only, no SMB)
- [4] Kolibri has landed (Aleph Alpha, vendor primary) - https://aleph-alpha.com/en/blog/kolibri-has-landed-a-sovereign-open-weight-model/ (checked 2026-10-03)
- [5] DeepKeep brings CISOs visibility into coding agents (PR Newswire) - https://www.prnewswire.com/news-releases/deepkeep-brings-cisos-visibility-into-coding-agents-and-ai-assisted-workflows-with-ai-lens-for-developers-302895883.html (wire, primary distribution, checked 2026-10-02)
- [6] Underdog founder note (vendor, primary) - https://underdog.im/founder (checked 2026-10-02)
- [7] Deepslate raises EUR 7.7M seed (The SaaS News) - https://www.thesaasnews.com/news/deepslate-raises-7-7m-seed/ (round details and investors, checked 2026-10-04)