Curiosity · AI Model
Claude Instant 1.2
Claude Instant 1.2 was Anthropic's early-2023 alternative to GPT-3.5 Turbo — a latency-optimised chat model with a 100K-token context window and a friendly, concise tone. It powered the first wave of Slack, Notion, and Quora integrations before being superseded by Claude 3 Haiku.
Model specs
- Vendor
- Anthropic
- Family
- Claude 1
- Released
- 2023-08
- Context window
- 100,000 tokens
- Modalities
- text
- Input price
- $0.8/M tok
- Output price
- $2.4/M tok
- Pricing as of
- 2024-06
Strengths
- Very fast responses relative to 2023-era peers
- 100K-token context was state-of-the-art at launch
- Low cost for the time
Limitations
- Deprecated — Anthropic recommends migration to Haiku 3.5 or newer
- No vision, tool use, or structured-output mode
- Reasoning quality well below 2025-era small models
Use cases
- Legacy chatbot endpoints with 100K-context summarisation
- Archived Slack/Notion integrations
- Reference baseline for cost/quality comparisons
Benchmarks
| Benchmark | Score | As of |
|---|---|---|
| MMLU | ≈74% | 2023-08 |
| HumanEval | ≈36% | 2023-08 |
Frequently asked questions
What is Claude Instant 1.2?
Claude Instant 1.2 is Anthropic's 2023 low-latency chat model, designed as a cheap alternative to Claude 1 and 2 with a 100K-token context window.
Is Claude Instant 1.2 still available?
Claude Instant 1.2 is deprecated. Anthropic has retired the endpoint for new customers and recommends migrating to Claude 3.5 Haiku or newer Haiku revisions.
What should I migrate to?
Claude Haiku 3.5 or Claude Haiku 4 are the recommended successors — cheaper, faster, and much stronger on reasoning and tool use.
Sources
- Anthropic — Model deprecations — accessed 2026-04-20
- Anthropic — Claude Instant announcement — accessed 2026-04-20