Anthropic cuts Haiku 5.5 pricing to match OpenAI’s GPT-6 Luna
Anthropic released Claude Haiku 5.5 on October 6, 2026, positioning it as a cheaper, faster alternative to OpenAI’s GPT-6 Luna at the same price. The model supports 1M-token context (up from 200K) and includes effort controls and adaptive thinking, but uses ~3x more tokens than Luna at max effort, offsetting some cost savings. Anthropic also cut Sonnet 5.5 cache-read costs by 50% and expanded…
Key points
- Anthropic priced Haiku 5.5 to match OpenAI’s GPT-6 Luna at $0.10/$0.50 per 1M tokens (under 100K tokens)
- Haiku 5.5 scores 43 on Intelligence Index, ahead of Luna (38) but behind Sonnet 5.5 (56) and Kimi K3 (44)
- Token usage is 3x higher than Luna at max effort, offsetting cost savings; partners report mixed but positive results
Independent benchmarks from Artificial Analysis show Haiku 5.5 scoring 43 on the Intelligence Index (up 26 points from Haiku 4.5), slightly ahead of Gemini 3.8 Flash (41) and GPT-6 Luna (38). It trails Sonnet 5.5 (56) but outperforms competitors like GLM-5.3 Flash (42) and Kimi K3 (44). However, token efficiency remains a concern: Haiku 5.5 uses ~162K tokens per task at max effort, compared to Luna’s ~50K. Partners like Cursor, GitHub Copilot, and Devin report cost savings and performance gains, though automation benchmarks are depressed by a pre-release safety bug. Anthropic has not yet published full cost-per-task figures for the new tiered pricing.
Model pages: GPT-6 Luna → · Sonnet 5.5 → · Kimi K3 →
The story so far
7 episodes →- Anthropic cuts Haiku 5.5 pricing to match OpenAI’s GPT-6 Lunathis story
[AINews] Claude Haiku 5.5 — better than GPT-6 Luna at the same pricing
Latent Space · 8 October 2026
Loading the full article…
This text was published by Latent Space and written by Latent.Space. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Google launches Gemini 4 Argon model · 13 src
- Meta introduces Muse Wonder and Muse Wonder Full models · 15 src
- Google upgrades SynthID to detect media from OpenAI and NVIDIA · 1 src
- Openbmb releases MiniCPM-V 4.7-35B-A3B, first open-source MoE model for on-device use · 2 src
- Jordan dismantles AI doom rhetoric as 'grand moralizing performance' · 1 src
Comments
via GitHub Discussions