Home / Directory / AI infrastructure / Inference & model hosting / Together AI
Together AI
Inference and fine-tuning platform for open models, with hosted endpoints and dedicated deployments.
Teams serving open models who want a hosted endpoint now and a dedicated option later.
Buyers standardized on a single closed-model vendor API, or shops that must run every model in their own VPC from day one.
Verdict
Together is a great fit for teams that need open-model inference and want someone else to run the serving stack at first. API pricing is competitive in the open-model host band, with a clear path to dedicated capacity once a model is a product dependency. Product teams comparing hosts on real prompts generally like the experience; pin the models you depend on as coverage changes.
Score Breakdown
How Together AI scores in the categories that matter to its buyers.
Pricing
Together AI publishes GPU Cluster and Dedicated Inference rates per GPU-hour alongside serverless per-token APIs. Representative on-demand USD GPU rates below are from Together’s public pricing page as of September 2026. Preemptible, reserved, and enterprise terms differ from these on-demand list rates.
| SKU | Product | On-demand / GPU / hr | What stands out |
|---|---|---|---|
| NVIDIA HGX H100 | GPU Clusters | $3.99 | Preemptible $1.99; reserved from $3.69 (7-30 days) |
| NVIDIA HGX H200 | GPU Clusters | $5.99 | Preemptible $2.99; reserved from $4.99 |
| NVIDIA HGX B200 | GPU Clusters | $8.19 | Preemptible $4.09; reserved from $7.99 |
| NVIDIA HGX H100 | Dedicated Inference | $3.99 | Promo through Sep 30, 2026; list $5.49 after |
| NVIDIA HGX B200 | Dedicated Inference | $8.99 | Reserved Dedicated Inference is sales quote |
| NVIDIA HGX H200 / B300 / GB200 | Dedicated Inference | Sales quote | Contact us on public card |
| Serverless chat models | Serverless Inference | Per 1M tokens | Example: Llama 3.3 70B $1.04 in / $1.04 out |
| Enterprise / long reserve | Clusters | Sales quote | 181+ day cluster reserve and GB200/B300 |
The Field at a Glance
How Together AI compares with other Inference & model hosting vendors we reviewed, by Overall Score and relative typical engagement cost.
Together AI scores 7.3 in Inference & model hosting, under Modal (8.4), Replicate (8.0), Beam (7.8), and Cerebrium (7.8). Relative cost lands upper for this subcategory.
Use-case matrix
| Use case | Fit | Notes |
|---|---|---|
| Hosted open-model inference | Strong | The main reason to start here. |
| Dedicated or reserved serving | Strong | The grown-up path once volume lands. |
| Closed-model only stack | Poor | Wrong catalog. |
| Full private cloud build | Mixed | Possible later; not the first-week product. |
Who it’s for
Good fit
- Product teams serving open models via API
- Groups comparing hosts on real prompt mixes
- Fine-tune users who also need inference
Poor fit
- Single-vendor closed-model shops
- Air-gapped environments on day one
- Buyers who will not benchmark their own traffic
Review Excerpts
Below are excerpts from public reviews. Paid reviews and pay-for-play sites such as Clutch were excluded.
“I’ve mostly settled on using a mixture of open weights models through Together.ai and Fireworks.ai, a MiniMax subscription for high-token-use tasks that don’t need the best model (for $20 I get what feels like infinite tokens), and codex for the occasional high complexity task, although with Kimi K3 and hopefully soon GLM 5.3, it’s becoming increasingly less important.”
“just tested (zai-org/GLM-5.3-Flash via together.ai) against latest DeepSeek-V4-Flash for a very specific task and thought i'd report here... - price: DS4 wins... $0.0235 vs $0.0242 for ten tasks - latency: GLM wins... 108s total against 154s this is for a personal use-case where i'm detecting ads in a written transcript. sticking with ds4-flash for now since latency is not a critical factor”
“Together serves models optimized for inference speed. They're not Groq but Together (and Perplexity Labs) have the lowest latencies and fastest tokens per second of any commercial services available right now. Also the lowest prices afaik.”
“We've been thoroughly impressed with Together. They delivered a 2x reduction in latency and cut our costs by approximately a third.”
“When you add a credit card it explicitly said: "There will be an immediate $1 charge for authorization, which will be credited back." I added my credit card, got charged $1, but never received it back.”
“After 2 weeks they still have not refunded the remaining balance on my prepaid account, nor would they refund the amount for troubleshooting the deactivated API key.”
Key Leadership
Vipul Ved Prakash
Co-founder and CEO
Methodology
This page is an independent evaluation of Together AI for buyers comparing options in inference & model hosting. AI Industry Reviews accepts no sponsorships, advertising, or pay-for-placement fees. Together AI did not pay for this review.
What we scored
The headline number is an Overall Score on a 0-10 scale. Eight criteria fall under it in two groups.
- Buyer outcomes (for inference & model hosting)
- Open-model endpoints
- Dedicated deployment path
- Fine-tune support
- Price and latency fit
- Company & commercial
- Innovation &
- product leadership
- Project management &
- communication
- Pricing
- Contract fairness
Pricing measures whether the price looks fair for the value delivered, including packaging and renewal friction that show up in real buying cycles.
Score Composition
| Input | Weight | What it covers |
|---|---|---|
| Reviews | 40% | A proprietary read of what practitioners say about likes, complaints, and day-to-day use, including public review sites, forums, and private chat rooms. Paid reviews and pay-for-play sites such as Clutch are out of scope. |
| Product | 35% | Hands-on look at screens and workflows. |
| Pricing | 15% | Whether the price looks fair for what you get. |
| Docs & training | 10% | Docs, tutorials, and training material. |
How we balanced the evidence
The Overall Score is the simple average of the eight criteria. Recommendation language follows that score and the fit pattern described above.
Scope
Together AI is graded here as inference & model hosting. Criteria scores can move as more review volume and product checks are added.