# Claude Opus 4.6-Level Performance Will Cost as Much as Claude 3.5 Haiku in 12 Months

Feb 7, 2026

**TL;DR:** Claude Opus 4.6 costs $10/M blended* today. Using HumanEval benchmark data and historical pricing trends, equivalent coding performance will cost $1.50-$2.00/M in 12 months—same as [Claude 3.5 Haiku](https://www.anthropic.com/claude/haiku) ($1.60/M) today.

* _Blended pricing assumes 3:1 input/output token ratio. Claude Opus 4.6 is $5 input / $25 output per 1M tokens._

* * *

## HumanEval: Price History

Writing Python functions from docstrings.

Price to achieve >65% HumanEval capability ($/1M tokens)

(Obviously the price won’t go to zero)

In March 2023, GPT-4 achieved 67% HumanEval for $37.50/M. By September 2024, Qwen2.5-Coder achieved 93% for $0.09/M. That’s a **417x price decline in 18 months** for far superior capability.

You can repeat the experiment with different benchmarks (GPQA Diamond, MMLU, etc), and it tells the same story.

### **Data sources** (HumanEval score, pricing at 3:1 input/output ratio)

- GPT-4 (Mar 2023): [67.0%](https://arxiv.org/abs/2303.08774), [$30/$60 = $37.50/M](https://openai.com/api/pricing/)
- Claude 2 (Jul 2023): [71.2%](https://www.anthropic.com/news/claude-2), [$8/$24 = $12.00/M](https://www.anthropic.com/news/claude-2)
- DeepSeek-Coder-33B (Nov 2023): [74.4%](https://deepwiki.com/deepseek-ai/DeepSeek-Coder/2.2-performance-and-benchmarks), [~$1.00/$2.00 = $1.25/M](https://deepwiki.com/deepseek-ai/DeepSeek-Coder/2.2-performance-and-benchmarks)
- Claude 3 Opus (Mar 2024): [84.9%](https://www.anthropic.com/news/claude-3-family), [$15/$75 = $30.00/M](https://www.anthropic.com/news/claude-3-family)
- GPT-4o (May 2024): [90.2%](https://openai.com/index/hello-gpt-4o/), [$2.50/$10 = $4.38/M](https://openai.com/api/pricing/)
- Claude 3.5 Sonnet (Jun 2024): [92.0%](https://www.anthropic.com/news/claude-3-5-sonnet), [$3/$15 = $6.00/M](https://www.anthropic.com/news/claude-3-5-sonnet)
- GPT-4o mini (Jul 2024): [87.2%](https://docsbot.ai/models/gpt-4o-mini), [$0.15/$0.60 = $0.26/M](https://docsbot.ai/models/gpt-4o-mini)
- o1-mini (Sep 2024): [92.4%](https://openai.com/index/openai-o1-mini-advancing-cost-efficient-reasoning/), [$3/$12 = $6.00/M](https://openai.com/api/pricing/)
- Qwen2.5-Coder-32B (Sep 2024): [92.7%](https://llm-stats.com/models/qwen-2.5-coder-32b-instruct), [$0.09/$0.09 = $0.09/M](https://llm-stats.com/models/qwen-2.5-coder-32b-instruct)
- Claude 3.5 Haiku (Nov 2024): [88.1%](https://docsbot.ai/models/claude-3-5-haiku), [$0.80/$4.00 = $1.60/M](https://docsbot.ai/models/claude-3-5-haiku)
- DeepSeek V3 (Dec 2024): [82.6%](https://llm-stats.com/models/deepseek-v3), [$0.27/$1.10 = $0.48/M](https://llm-stats.com/models/deepseek-v3)

Our users keep asking for more usage. This is why I tell them to expect 10x cheaper usage in a year.

* * *