C
Claude Haiku 5.5
The fastest lightweight model here (235 tok/s), close to Opus on LiveBench design tasks
Intelligence v5
36.2
Leaderboard rank
#11
Input / output
$0.10/$0.50
Blended 3:1
$0.20
Output speed
235 tok/s
Value
181 pts/$
Capability profile Normalised to the field maximum
Specifications
Vendor
Anthropic
Released
2026-09
Weights
Closed, API only
Context
1,000,000 tokens
Max output
128,000 tokens
Parameters
Undisclosed
Input price
$0.10 / 1M
Output price
$0.50 / 1M
Cache hit
$0.01 / 1M
Output speed
235 tok/s
Intelligence
36.2（v5）
Benchmark scores
How it is measured →Ranked 11 · bars are scaled to the field maximum for that benchmark.
| Benchmarks | Score | Vs. best in field | Gap |
|---|---|---|---|
| AA Intelligence Index v5 Intelligence index | 36.2 | 61% | −23.2 pts |
| AA Intelligence Index v4.3.2 Intelligence index (legacy) | 43.0 | 75% | −14.6 pts |
| GPQA Diamond | — | — | not indexed |
| Terminal-Bench 4.0 | — | — | not indexed |
| LiveBench Design Document generation (DOCX/PPT/PDF) | 83.1 | 96% | −3.8 pts |
| Vals Index | — | — | not indexed |
Assessment
A LiveBench design score of 83.1, only 3.8 points behind Opus 5.5, at 1/40 of the price. Note that $0.10/$0.50 only applies to prompts under 100K tokens; beyond that it bills at $0.50/$2.50.
At the same price
The four models closest on blended price, side by side.
| Model | v5 | Blended | tok/s | Score gap |
|---|---|---|---|---|
| MiniMax M3.1 FlashMiniMax | 29.0 | $0.26 | 118 | -7.2 |
| DeepSeek V4.1 Flash Open weightsDeepSeek | 32.5 | $0.53 | 148 | -3.7 |
| MiMo-V2.6 Pro Open weightsXiaomi | 30.0 | $0.61 | 81 | -6.2 |
| GLM-5.3 Open weightsZ.ai | 32.0 | $1.00 | 77 | -4.2 |
Compare side by side
Line it up against other models on price, speed and every benchmark.
The comparison tool lines up to 4 models side by side with switchable normalisation — built for purchasing decisions.
Open the comparison tool → Back to leaderboard