When you need to score reviews, classify support tickets, summarize emails at scale — Claude API costs add up. DeepSeek matches frontier reasoning on most batch tasks at ~10x cheaper. Use Claude to write the prompt and the eval set; switch to DeepSeek for production volume; sample-check with Claude monthly.
Use Claude to draft the prompt, hand-label 30 to 50 golden examples, and define what 'correct' looks like for your task. This is the work where model quality matters most.
DeepSeek-V3 / R1 hit ~$0.14/M input + $0.28/M output. On most classification, summarization, and structured-extraction tasks, output quality is within 2 to 5% of Claude on the eval set. The 90%+ cost savings dominate at volume.
Fast mode for Claude Opus 4.8 is being deprecated and will be removed on November 30, 2026. Developers using fast mode with Opus 4.8 must migrate to Claude Opus 5.5 to continue using the feature.
Anthropic released the Usage and Cost Analytics API in beta, enabling organizations to monitor token usage and costs across Claude API consumption with filtering and grouping by dimensions like model, workspace, and API key. The API provides /v1/organizations/analytics/usage and /v1/organizations/analytics/cost endpoints. Legacy usage reporting endpoints remain supported.
The maximum request size for the Messages API and Token Counting API doubled from 32 MB to 128 MB on the first-party Claude API, allowing developers to process significantly larger payloads in single requests. This change takes effect immediately and applies to both APIs.
The Claude Tag API is now available in beta for Claude Enterprise organizations, providing endpoints to read and configure Claude Tag settings on Slack channels and workspaces with Access bundle support. The API requires Admin API keys with read:claude_tag or write:claude_tag scope. This feature is exclusive to Claude Enterprise.
Use Claude's free tier — for prompt iteration, the chat surface is fine. The model quality matters here.
I'm building a production LLM task. Help me draft the prompt and an eval set.
Task description:
"""
{{describe the task: input, expected output format, what 'correct' means}}
"""
Output, in this order:
1. **System prompt** — the production prompt that will run on every input. Strict output format. No conversational hedging.
2. **30-row eval set** — table of {input, expected output, why}. Cover the easy cases, the edge cases, and 5 deliberately adversarial inputs.
3. **Eval rubric** — exactly how I'll score outputs against the expected outputs. Define partial credit if useful.
4. **Smoke test plan** — the 3 minimum-viable runs I should do before sending real volume to a cheaper model.
The point of all this: I'm going to run the prompt on 100k inputs/month via DeepSeek's API. I want a prompt that survives the swap from Claude (where I'm authoring it) to DeepSeek (where it'll run).Run the eval set through both Claude and DeepSeek. If DeepSeek scores within 5% of Claude on your rubric, ship DeepSeek. If not, narrow the prompt or the task scope before re-testing.
Standard OpenAI-compatible API. Most SDKs work by changing the base URL.
Sample 1 to 2% of DeepSeek outputs and re-score with Claude. Catches drift and prompt rot before they become a customer-facing problem.
Classifying ~120k tickets/month into 14 categories. Claude API: ~$2,800/mo at this volume. DeepSeek: ~$210/mo. Eval-set quality: within 3.4% on F1 (Claude 91.2 vs DeepSeek 87.8). Saved ~$31k/year; spent the savings on better hand-labeling.
The whole bet of this stack is that DeepSeek matches Claude on YOUR task. Without an eval set, you're guessing. The eval set is the cheapest insurance possible.
Production LLMs all hallucinate sometimes. Don't switch on n=1 — the right move is to widen the eval set and confirm a real quality gap before paying 10x more.
It's offered as a hosted API, but the model weights are open. If data residency or supplier risk is a concern, you can self-host the same model on your own infra. Most teams won't, but knowing the option exists changes the conversation.
What changed in AI tools and the stacks that ship. One email a week.