Skip to main content
idapt
HomeCodeAI ModelsPricing
Sign inStart free
  • Home
  • Pricing
  • AI Models
  • Image models
  • Voice models
  • Video models
  • Rankings
  • New models
  • Model status
  • Multi-Model Chat
  • Agents
  • Computers
  • Drive
  • Automations
  • AI Gateway
  • All features →
  • LLM cost calculator
  • Token counter
  • All free tools →
  • Blog
  • Use cases
  • Comparisons
  • Best of
  • Benchmarks
  • Demos
  • Changelog
  • Help center
  • FAQ
  • Privacy
  • Compare all models
  • Support
  • idapt Code
  • Developers
  • Quickstarts
  • API reference
  • API pricing
  • CLI
  • MCP
  • Downloads
  • Desktop
  • Badges and embeds
© idapt[email protected]TermsPrivacy PolicyLegal noticeReport content
X (Twitter)

GPT 5.6 Luna vs Gemini 3.8 Flash vs Claude Haiku 4.5

Fast models carry most day-to-day work. These three trade a little peak capability for speed and price; the columns show exactly how much of each.Last reviewed 2026-09-09.

GPT 5.6 Luna vs Gemini 3.8 Flash vs Claude Haiku 4.5

At a glance

OpenAI's fastest, most affordable GPT-5.6 tier for high-volume, latency-sensitive work.

1.1M ctx$1.00/$6.00 per M

Google's most intelligent Flash model, with significant gains over 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning.

1.0M ctx$0.75/$3.75 per M

Fast and efficient with near-frontier intelligence.

75200K ctx$1.00/$5.00 per M#41/92
Demos
All demos

No shared demos for these models yet.

AI Output
All AI outputs

No captured outputs for these models yet.

Benchmarks
See rankings
The paired significance test (McNemar, in "Measured on Idapt") appears only when exactly two models are pinned.
Reasoning
CodingInsufficient data
MathInsufficient data
Sources:Provider·OpenRouter· as of 2026-07-11
Capability
Capability—
Reasoning & Knowledge
Graduate Science53%92.3%
Math
Math Problems—Competition Math—Research Math—
Agentic
Terminal Tasks184%84.7%Novel Reasoning—
Reasoning
CodingInsufficient data
AgenticInsufficient data
ReasoningInsufficient data
MathInsufficient data
Sources:OpenRouter
Cheaper25%
Capability
Capability—
Reasoning & Knowledge
Graduate Science—
Math
Math Problems—Competition Math—Research Math—
Agentic
Terminal Tasks—Novel Reasoning—
Reasoning
CodingInsufficient data
Sources:Epoch ECI·Epoch AI·OpenRouter· as of 2026-07-10
Capability
CapabilityECI75.3
Reasoning & Knowledge
Graduate Science60.5%
Math
Math Problems86.9%Competition Math35.8%Research Math4.1%
Agentic
Terminal Tasks29.8%Novel Reasoning1.3%
Specs
1.1M
Context
128K
Max output
Jul 2026
Released
1.0M
Context
66K
Max output
Sep 2026
Released
200K
Context
64K
Max output
Oct 2025
Released
Capabilities
Vision
Audio
Reasoning
Vision
Audio
Reasoning
Vision
Audio
Reasoning
Pricing
Live pricing, shown before you run
Input$1.00/M tokens
Output$6.00/M tokens
Cache read$0.10/M tokens
Cache write$1.25/M tokens
Web search$0.01/search
Live pricing, shown before you run
Input25%$0.75/M tokens
Output25%$3.75/M tokens
Reasoning$3.75/M tokens
Cache read25%$0.07/M tokens
Cache write97%$0.04/M tokens
Image input$0.75/M tokens
Audio input$0.75/M tokens
Web search$0.01/search
Live pricing, shown before you run
Input$1.00/M tokens
Output$5.00/M tokens
Cache read$0.10/M tokens
Cache write$1.25/M tokens
Web search$0.01/search
Chat with GPT 5.6 LunaGo to model
Chat with Gemini 3.8 FlashGo to model
Chat with Claude Haiku 4.5Go to model

Frequently asked

Which of GPT 5.6 Luna, Gemini 3.8 Flash, and Claude Haiku 4.5 is the cheapest?

Gemini 3.8 Flash has the lowest blended list price. Per 1M tokens: GPT 5.6 Luna at $1.00 in / $6.00 out; Gemini 3.8 Flash at $0.75 in / $3.75 out; Claude Haiku 4.5 at $1.00 in / $5.00 out.

Which of GPT 5.6 Luna, Gemini 3.8 Flash, and Claude Haiku 4.5 has the largest context window?

GPT 5.6 Luna leads with 1.1M tokens. The other two: Gemini 3.8 Flash at 1M, Claude Haiku 4.5 at 200K.

Do GPT 5.6 Luna, Gemini 3.8 Flash, and Claude Haiku 4.5 support image input?

Yes, all three accept images.

Where can I run GPT 5.6 Luna, Gemini 3.8 Flash, and Claude Haiku 4.5 side by side?

This idapt comparison runs all three in aligned columns: the same prompt, plus benchmarks, specs, and prices per model. Open any model in chat from its column to keep working with it.

More three-way comparisons

GPT 6 Astra vs Claude Opus 5 vs Gemini 3.1 ProClaude Fable 5.1 vs GPT 6 Astra vs Grok 4.6Claude Sonnet 5 vs Kimi K2.7 Code vs Qwen3 Coder NextDeepSeek V4 Pro (0813) vs Qwen3.8 2.4T A95B vs Nemotron 3 UltraGLM 5.3 Flash vs MiniMax M3 vs DeepSeek V4 FlashClaude Sonnet 5 vs Grok Build 0.1 vs Kimi K3Muse Spark 1.3 vs Kimi K3 vs Grok 4.6DeepSeek V4 Flash vs Gemini 3.8 Flash vs Qwen3.7 PlusGLM 5.3 vs MiniMax M3 vs DeepSeek V4 Pro (0813)