Pairwise benchmark snapshot · Aug 1, 2026
In this head-to-head showdown, GPT-5 delivers higher overall intelligence and human-preferred responses (Elo 1,558 vs 1,224), while GPT-5 leads in SWE-bench software engineering benchmarks (68.4%), while Claude 3.5 Haiku is more budget-friendly at $4/1M per 1M output tokens. Review the complete breakdown below to determine which model best fits your performance and budget requirements.
9 direct benchmark disciplines evaluated across capability, speed, and cost
Previous cheap Claude. Haiku 4.5 is the current $1/$5 SKU.
OpenAI flagship reasoning model for 2025–26. Strong general preference Elo and multimodal coverage.
Simulate accuracy gains vs added response time & token cost
Reasoning models “think before answering” by generating internal reasoning tokens. Higher effort improves math, coding, and logic accuracy, but increases response delay and token costs.
Δ 334 Arena Elo pts
140 tok/s
$4 / 1M output
Relative percentile scores computed across all active models in the benchmark catalog.
Comparing top Western standard models with China's leading frontier rival across 6 skill dimensions. Tap any spoke or dot to inspect.
Category wins across reasoning intelligence, generation speed, and token cost.
Preference Elo, Coding proficiency, SWE-bench & LiveBench accuracy
Generation throughput and time to first token responsiveness
Cost per million tokens and max context window length
| Benchmark | Claude 3.5 Haiku | GPT-5 | Advantage Delta |
|---|---|---|---|
| Preference Elo | +334 | ||
| Coding Elo | — | — | |
| LiveBench | +25% | ||
| SWE-bench | +32.2% | ||
| GPQA Diamond | +31.1% | ||
| Time to first token | +180 ms | ||
| Output speed | 140 tok/s | 78 tok/s | +62 tok/s |
| Input price | $0.8/1M | +$4.2/1M | |
| Output price | +$16/1M | ||
| Context window | +200k |
Simulate monthly production API costs in USD (US Dollar).
Claude 3.5 Haiku is estimated to save $387.00/month ($4,644/year).
Higher SWE-bench (68.4%).
Lower output list price ($4/1M).
Lower TTFT (110 ms).
Larger window (400k).
Both accept images. Defaulting to the higher-Elo side (GPT-5).
| Target Workload | Recommended Pick | Evaluation Rationale |
|---|---|---|
| Repo / coding agents | GPT-5 | Higher SWE-bench (68.4%). |
| High-volume chat | Claude 3.5 Haiku | Lower output list price ($4/1M). |
| Voice / low-latency UI | Claude 3.5 Haiku | Lower TTFT (110 ms). |
| Long-document RAG | GPT-5 | Larger window (400k). |
| Screenshots / vision | GPT-5 | Both accept images. Defaulting to the higher-Elo side (GPT-5). |
No comments posted on this matchup yet. Be the first to share an evaluation note!
Dynamic OpenGraph banner generated at /compare/claude-3-5-haiku-vs-gpt-5/opengraph-image
Verified Head-to-Head Telemetry