Claude Haiku 5.5 vs GPT-6 Luna

$0.047 / task
Wins 1 of 4 benchmarks
$0.017 / task
Wins 0 of 4 benchmarks
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Haiku 5.5—
GPT-6 Luna77.9
GDP.pdf · AA all-pass GDP.pdf · AA all-pass: Models reason over professional PDF documents; an attempt passes only when every atomic criterion passes
Haiku 5.515.2
GPT-6 Luna13.8
Chartography Chartography: Expert chart reasoning over professional scientific, medical, financial and engineering charts
Haiku 5.5—
GPT-6 Luna · max29.1
Arena Vision · overall Arena Vision · overall: Human pairwise preferences in the stated vision vote pool
Haiku 5.5—
GPT-6 Luna · max1,216.1

Specs

SpecHaiku 5.5GPT-6 Luna
Input, per 1M$0.10$0.10
Output, per 1M$0.50$0.50
Cost per task?$0.047$0.017
Speed, tok/s179—
Context1M1.05M
WeightsClosedClosed
ReleasedOct 2026Sep 2026
See images✓✓
Read PDFsPartialPartial
Hear audio——
Watch video——
Call tools✓✓
Force a tool callPartialPartial
Return strict JSON✓✓
Search the web✓✓
Run code✓✓
Use a computer?✓

What breaks when you switch

Haiku 5.5 → GPT-6 Luna
Function calling only with none · Chat Completions›

Chat Completions function calling requires reasoning_effort none; use Responses for reasoning with tools.

More than 272K input tokens›

Prompts above 272K input tokens charge 2× input and cache rates and 1.5× output rates for the entire request.

Remove unsupported sampling parameters›

When reasoning effort is not none, remove temperature, top_p and top_logprobs, plus Chat Completions logprobs or Responses output logprobs inclusion.

Cache writes billed at 1.25x›

Cache writes cost 1.25× the uncached input token rate.

GPT-6 Luna → Haiku 5.5
Context: 1.05M → 1M
Remove sampling parameters›

Unsupported sampling values, any top_k or combined temperature and top_p return HTTP 400; omit all three and use prompting.

Replace manual thinking budgets›

Manual thinking with budget_tokens returns HTTP 400; replace the old thinking configuration.

No assistant prefill›

A final assistant prefill returns HTTP 400 even with thinking off; end messages with a user turn.

Replace old computer tool · Claude API / Google Cloud›

computer_20250124 returns HTTP 400 on the Claude API and Google Cloud; use computer_toolset_20260801 there.

Keep conversations append-only›

Replaying thinking after earlier history changes returns HTTP 400, with older accounts checked only when the binding field is set; keep history append-only.

Bedrock needs another computer version · Bedrock›

Bedrock does not accept computer_20250124; use computer_20251124 with the computer-use-2025-11-24 beta header.