Together AI overhauls AI catalog with 3 new models, cuts legacy lineup
What happened
Between the Feb 9, 2026 and Feb 18, 2026 captures of Together AI's pricing page, Pulse recorded 5 changes: Removed legacy Llama models from Serverless Inference catalog: Llama 3 70B Instruct Reference ($0.88/$0.88 per 1M tokens) and LLaMA-2 ($0.90/$0.90 per 1M tokens); Removed M2-BERT 80M 32K Retrieval ($0.01/1M tokens) embedding model and Salesforce LlamaRank ($0.10/1M tokens) reranking model from catalog; Added 3 new text models: MiniMax M2.5 ($0.30/$1.20 per 1M tokens), Qwen3.5-397B-A17B ($0.60/$3.60 per 1M tokens), and GLM-5 ($1.00/$3.20 per 1M tokens) to Serverless Inference catalog; Removed third-party model families from Serverless Inference: Mistral Instruct ($0.20/$0.20), Arcee AI Coder-Large/Maestro/Virtuoso-Large, Cogito v2 preview series (109B/405B/671B/70B), and Refuel LLM-2/LLM-2 Small; Removed Qwen2.5 model family from Serverless Inference catalog: Qwen3 235B A22B FP8 Throughput ($0.20/$0.60), Qwen2.5 72B ($1.20/$1.20), Qwen2.5 Coder 32B Instruct ($0.80/$0.80), Qwen2.5-VL 72B Instruct ($1.95/$8), and Qwen QwQ-32B ($1.20/$1.20). Pulse files it as a product change (feature change) with a neutral direction. Pulse's latest capture (Oct 21, 2026) lists Managed Storage at $0.16 per GiB/month, Provisioned Throughput at $0.05 per PTU/minute, Sandbox (Code Interpreter) at $0.03 per session (60 min), GPU Clusters (On-demand) at $3.99 per hour (H100) and GPU Clusters (Reserved) at $3.69 per hour (H100, 3-30 days).
Together AI's previous recorded change was "Together AI hikes DeepSeek-V3.1 output price, slashes Mistral Small 3 rates" (Week 7, 2026). Before that Pulse recorded "Together AI adjusts pricing on key models, with notable decreases and increases" (Feb 10, 2026) and "Together AI cuts DeepSeek-V3.1 prices 26%, raises GPU cluster rates" (Week 3, 2026). Together AI has been tracked since 2025-12-23.
In the same window Anthropic, OpenAI and Huggingface also changed their AI & Machine Learning pricing pages. Those moves are listed further down, and the Pulse Agent can compare them side by side.
- Serverless Inference · FeatureRemovedlegacy Llama modelsNo legacy Llama models
- — · FeatureRemoved
BeforeM2-BERT 80M 32K Retrieval
AfterNo M2-BERT 80M 32K Retrieval
- Serverless Inference · FeatureAddedNo new text models3 new text models
- Serverless Inference · FeatureRemoved
Beforethird-party model families
AfterNo third-party model families
- Serverless Inference · FeatureRemovedQwen2.5 model familyNo Qwen2.5 model family
| Plan | Field | Before | After | Change |
|---|---|---|---|---|
| Serverless Inference | Feature | legacy Llama models | No legacy Llama models | Removed |
| — | Feature | M2-BERT 80M 32K Retrieval | No M2-BERT 80M 32K Retrieval | Removed |
| Serverless Inference | Feature | No new text models | 3 new text models | Added |
| Serverless Inference | Feature | third-party model families | No third-party model families | Removed |
| Serverless Inference | Feature | Qwen2.5 model family | No Qwen2.5 model family | Removed |
Values as written on Together AI's public pricing page in the two captures Pulse compared; where the page states no number, the cells give the wording before and after.
Evidence
The two captures Pulse comparedFull-page captures of Together AI's pricing page (www.together.ai) from Feb 9, 2026 and Feb 18, 2026; the change is marked near the top of the page, where the legacy Llama models no longer appears.
Sign in to see the evidence
The two full-page captures with this change marked up are free with an account.
Together AI, the last 90 days
Full history- Apr 3, 2026packagingTogether AI Raises GPU Prices Up to 36% and Restructures Cloud Tiers
- Mar 17, 2026packagingTogether AI Adds Qwen3.5 9B, Drops Five Models in Serverless Lineup Shakeup
- Mar 11, 2026packagingTogether AI adds Llama 4 Maverick at $0.85, revamps GPU pricing
- Feb 18, 2026improvedTogether AI overhauls AI catalog with 3 new models, cuts legacy lineup · this change
- Week 7, 2026increaseTogether AI hikes DeepSeek-V3.1 output price, slashes Mistral Small 3 rates
- Feb 10, 2026increaseTogether AI adjusts pricing on key models, with notable decreases and increases.
- Week 3, 2026increaseTogether AI cuts DeepSeek-V3.1 prices 26%, raises GPU cluster rates
Across AI & Machine Learning in the same window
All AI & Machine Learning changesQuestions about this change
Did Together AI raise prices in February 2026?
No list-price increase was recorded between Feb 9, 2026 and Feb 18, 2026. The change Pulse detected was a product change: feature change.
What changed on Together AI's pricing page between Feb 9, 2026 and Feb 18, 2026?
Serverless Inference feature: legacy Llama models → No legacy Llama models, removed; feature: M2-BERT 80M 32K Retrieval → No M2-BERT 80M 32K Retrieval, removed; Serverless Inference feature: No new text models → 3 new text models, added; Serverless Inference feature: third-party model families → No third-party model families, removed; Serverless Inference feature: Qwen2.5 model family → No Qwen2.5 model family, removed. Pulse compared the captures it took on Feb 9, 2026 and Feb 18, 2026; the before/after table above lists each item as written on the page.
How often does Together AI change its pricing?
Pulse recorded 3 other pricing-page changes for Together AI in the 90 days before this one; the previous change was "Together AI hikes DeepSeek-V3.1 output price, slashes Mistral Small 3 rates" in Week 7, 2026; Together AI has been tracked since 2025-12-23.