Which model is better: Claude 3.5 Haiku or Gemini 2.5 Pro?
Claude 3.5 Haiku is 36.0% cheaper for input tokens ($0.80 vs. $1.25 per 1M tokens) and $4.00 vs. $10.00 for output tokens (1.6x cost advantage). In terms of operational performance, Claude 3.5 Haiku delivers lower response latency at 190ms TTFT (510ms faster than Gemini 2.5 Pro) with 130 vs. 80 tok/s throughput. Gemini 2.5 Pro leads coding evaluations at 63.8% SWE-bench Verified vs. 40.6% for Claude 3.5 Haiku and 74.0% vs. 41.6% on GPQA Diamond.
Verified daily via automated API latency tests and official documentation.
Zero-Shot RAG Verified
Verified Performance & Unit Economics Delta (12 Dimensions)
Daily synchronized metrics across token pricing, prompt cache economics, latency, throughput, and 5 capability benchmarks.
Metric / Dimension
Claude 3.5 Haiku
Gemini 2.5 Pro
Net Delta / Advantage
Input Cost / 1M Tokens
$0.80
$1.25
Claude 3.5 Haiku is 1.6x cheaper
Output Cost / 1M Tokens
$4.00
$10.00
Claude 3.5 Haiku (60.0% lower)
Prompt Cache Read / 1M
$0.080
$0.313
Claude 3.5 Haiku is 3.9x cheaper
Prompt Cache Write / 1M
$1.00
$1.25
Claude 3.5 Haiku is 1.3x cheaper
Context Window (Max Output)
205k (8k out)
1,024k (66k out)
Gemini 2.5 Pro (5.0x larger context)
Avg TTFT Response Latency
190ms
700ms
Claude 3.5 Haiku (510ms faster)
Generation Speed (Throughput)
130 tok/s
80 tok/s
Claude 3.5 Haiku (+50 tok/s faster)
SWE-bench Verified (Coding)
40.6%
63.8%
Gemini 2.5 Pro (+23.2% lead)
GPQA Diamond (PhD STEM)
41.6%
74%
Gemini 2.5 Pro (+32.4% lead)
MMLU-Pro (Multi-Discipline)
71%
85.2%
Gemini 2.5 Pro (+14.2% lead)
MATH-500 (Quantitative Logic)
69.4%
95%
Gemini 2.5 Pro (+25.6% lead)
Inference Value Score (IVR)
52.4 / 100
56.8 / 100
Gemini 2.5 Pro (+4.4 pts)
Interactive Monthly Token Economics & ROI Forecaster
Model production workloads across prompt, completion, and cached context tokens, or switch models on the fly.
Choose Gemini 2.5 Pro (Google) for long-context document and video analysis, large codebase ingestion, and multimodal reasoning — offering a 1,024k context window, 700ms TTFT, and $1.25/1M input pricing ($0.312/1M cached reads).
Which is cheaper: Claude 3.5 Haiku or Gemini 2.5 Pro?
Claude 3.5 Haiku is cheaper at $0.80/1M input tokens ($4.00/1M output) compared to $1.25/1M input ($10.00/1M output) for Gemini 2.5 Pro.
Which model has lower TTFT latency and higher tok/s throughput: Claude 3.5 Haiku or Gemini 2.5 Pro?
Claude 3.5 Haiku delivers faster first-token latency at 190ms TTFT (130 tok/s throughput) compared to 700ms TTFT (80 tok/s) for Gemini 2.5 Pro.
How do Claude 3.5 Haiku and Gemini 2.5 Pro compare on SWE-bench coding and GPQA Diamond reasoning?
On SWE-bench Verified software engineering, Claude 3.5 Haiku scores 40.6% versus 63.8% for Gemini 2.5 Pro. On GPQA Diamond graduate-level STEM reasoning, Claude 3.5 Haiku scores 41.6% versus 74% for Gemini 2.5 Pro (and 69.4% vs 95% on MATH-500).
When should you choose Claude 3.5 Haiku?
Choose Claude 3.5 Haiku (Anthropic) for rapid customer-facing chat, sub-agent routing, and real-time code autocomplete — offering a 204k context window, 190ms TTFT, and $0.80/1M input pricing ($0.080/1M cached reads).
When should you choose Gemini 2.5 Pro?
Choose Gemini 2.5 Pro (Google) for long-context document and video analysis, large codebase ingestion, and multimodal reasoning — offering a 1,024k context window, 700ms TTFT, and $1.25/1M input pricing ($0.312/1M cached reads).
