Claude Haiku 5.5

AnthropicProprietaryReleased Oct 7, 2026

On the arena.ai Agents leaderboard, Claude Haiku 5.5 has a Net Improvement of 0.62% (High, rank #28, 95% interval -0.40%–1.65%). Claude Haiku 5.5 ranks #128 on the LMArena Text leaderboard (style control) with an Arena score of 1423.2 (±6.4, 12,481 votes). Artificial Analysis Intelligence Index: 36.5. API price: $0.10 input / $0.50 output per 1M tokens. Context: 1M.

Quick answer

Agents leaderboard
#28 of 39 · 0.62% success
Text (LMArena)
#128 · 1423.2
17NAS overall
#23 of 38
API price (in / out)
$0.10 / $0.50 per 1M tokens
Closest rival
Qwen3 Max · 1423.5
Data updated
Nov 7, 2026 · synced daily

Sources: arena.ai, LMArena, Artificial Analysis, OpenRouter

Key metrics

Agent Net Improvement0.62%High31,204 sessions · #28
Arena score1423.2±6.4LMArena #128 · 12,481 votes
Intelligence36.5Variant: max
Coding–Artificial Analysis
Writing (creative)1336.8±12.4LMArena creative writing #186 / 417
List price in / out$0.10 / $0.50USD per 1M tokens
Lowest route$0.10 / $0.50Cheapest OpenRouter route
Context1Mtokens
Output speed234tok/sAA measured median
TTFT / first answer236 / 236sReasoning model: includes thinking

Rank history

LMArena rank, last 16 days
#125#127#129Oct 23Nov 7

#126 on Oct 23, #128 now; best #126 in this period.

Reasoning effort comparison

Artificial Analysis measurements per reasoning effort of the same model
EffortIntelligenceCodingtok/sTTFT (s)First answer (s)
maxUsed in table 36.5–234236236
xhigh 34.6–1735050
high 29.9–1441616
medium 24.6–1447.87.8
low 18.6–1394.64.6

Models ranked nearby

Sorted by Arena score
#ModelArenaIntel.Output $tok/s
61DeepSeek V3.21424.7–$0.42–
62GLM-4.61424.2–$2.20–
63Qwen3 Max1423.5–––
64Claude Haiku 5.5This page1423.236.5$0.50234

More from Anthropic

#ModelArenaIntel.Output $tok/s
4Claude Opus 5.51505.159.1$20.0096
5Claude Fable 51504.0–$50.00–
6Claude Fable 5.11503.749.2$50.0068
7Claude Opus 4.61503.2–$25.00–
8Claude Opus 4.71501.0–$25.00–
12Claude Opus 51490.4–$25.00–

Pricing

  • Official API list price: $0.10 input and $0.50 output per 1M tokens (USD).
  • At a typical 3:1 input-to-output ratio, the blended cost is about $0.20 per 1M tokens.
  • Among 68 models with an output price, it is the #4 cheapest.
  • The cheapest OpenRouter route charges $0.10 input / $0.50 output.

Sources & methodology

On the arena.ai Agents leaderboard, Claude Haiku 5.5 has a Net Improvement of 0.62% (High, rank #28, 95% interval -0.40%–1.65%).

  • arena.ai Agents — agent Net Improvement, 95% interval, sessions and rank (via the public Hugging Face dataset, CC BY 4.0).
  • LMArena — Arena score (Text leaderboard with style control, the same view arena.ai shows by default), 95% confidence interval, votes and rank.
  • Artificial Analysis — Intelligence / Coding Index, output speed and latency (medians).
  • OpenRouter — Route prices, context length and listing date.

Arena scores use the Bradley-Terry model (not Elo); ± is the half-width of the 95% confidence interval. Latency figures are Artificial Analysis medians; for reasoning models, time to first token and to first answer include the thinking phase. Data syncs daily at 14:20 (UTC+8).

Limits: Arena scores come from anonymous users voting on chats, so they reflect conversational preference rather than agent or coding benchmarks; with few votes the interval is wide and ranks move. Prices and speeds change as providers update them.

FAQ

How does Claude Haiku 5.5 do on the Agents leaderboard?

On the arena.ai Agents leaderboard, Claude Haiku 5.5 has a Net Improvement of 0.62% (High, rank #28, 95% interval -0.40%–1.65%).

What rank is Claude Haiku 5.5 right now?

Claude Haiku 5.5 is #64 on our leaderboard of 64 scored models (Arena score 1423.2); its raw LMArena rank is #128.

What does Claude Haiku 5.5 cost via API?

Input $0.10 / output $0.50 per 1M tokens (official API list price; the OpenRouter quote when no official price is available). At a typical 3:1 input-to-output ratio, the blended cost is about $0.20 per 1M tokens.

How fast is Claude Haiku 5.5?

Artificial Analysis measures about 234 output tokens per second, 236 s to the first token and 236 s to the first answer token; it is a reasoning model, so latency includes thinking time.

Is Claude Haiku 5.5 open source? Can it be self-hosted?

Claude Haiku 5.5 is a closed model, available only via official APIs or hosted services.

What is the context length of Claude Haiku 5.5?

1M (1,000,000 tokens, as listed on OpenRouter).

Claude Haiku 5.5 vs Qwen3 Max: which is better?

On the LMArena Text leaderboard Claude Haiku 5.5 scores 1423.2 and Qwen3 Max scores 1423.5 (a 0.3-point gap, within the confidence interval, so not significant). On price, Claude Haiku 5.5 costs $0.50 and Qwen3 Max – per 1M output tokens.

Is there a cheaper model close to Claude Haiku 5.5?

Yes: DeepSeek V4 Flash scores 1435.7 on the Text leaderboard (Claude Haiku 5.5: 1423.2) and costs $0.28 per 1M output tokens, about 44% less than Claude Haiku 5.5 ($0.50).

Where does the Claude Haiku 5.5 data come from and how often is it updated?

Agent success rates come from the arena.ai Agents leaderboard (CC BY 4.0), Text scores from LMArena, intelligence and speed from Artificial Analysis, and prices and context from official list prices and OpenRouter. 17NAS syncs daily at 06:20 UTC; this page was last updated Nov 7, 2026.

Related leaderboards