What the vendor claims, next to what independent boards measured. The two are never mixed in one column.
Source: OpenRouter, as of Oct 28, 09:19 UTC. Model page on OpenRouter
Quoted word for word from the vendor's own launch post or model card. Claims, not measurements.
No vendor claims recorded yet. They come from launch posts and model cards the Drops covered.
Measured by independent sources (Epoch AI, LMArena, OpenRouter's benchmark runs, τ²-bench, Vectara), not by the vendor. "Listed as" is the source's own name for the model, so you can check our match.
| Board | Result | Rank |
|---|---|---|
| Epoch AI · FrontierMath-Tiers-1-3-v2-PrivateListed as: claude-haiku-5-5_maxSource: Epoch AI (CC BY), as of Oct 28, 2026. Source | 75.1%Accuracy | 21 |
| Epoch AI · FrontierMath-Tier-4-v2-PrivateListed as: claude-haiku-5-5_maxSource: Epoch AI (CC BY), as of Oct 28, 2026. Source | 46.3%Accuracy | 28 |
| Epoch AI · OTIS Mock AIME 2024-2025Listed as: claude-haiku-5-5_xhighSource: Epoch AI (CC BY), as of Oct 28, 2026. Source | 98.9%Accuracy | 21 |
| Epoch AI · Chess PuzzlesListed as: claude-haiku-5-5_maxSource: Epoch AI (CC BY), as of Oct 28, 2026. Source | 55.0%Accuracy | 16 |
| Epoch AI · Furniture AssemblyListed as: claude-haiku-5-5_maxSource: Epoch AI (CC BY), as of Oct 28, 2026. Source | 47.5%Accuracy | 13 |
| Epoch AI · Epoch Capabilities IndexListed as: Claude Haiku 5.5Source: Epoch AI (CC BY), as of Oct 28, 2026. Source | 159.32Index | 16 |
| LMArena · AgentListed as: Claude Haiku 5.5 (High)LMArena (CC BY 4.0), as of Oct 23, 2026. Source | 0.0114Score | 24 |
| OpenRouter · τ²-bench airlineListed as: Claude Haiku 5.5via Anthropic, its own providerSource: OpenRouter (openrouter.ai), licensed under CC BY 4.0, as of Oct 28, 2026. Source | 74.2%Accuracy | 58 |
| OpenRouter · τ²-bench airline: cost per taskListed as: Claude Haiku 5.5via Anthropic, its own providerLower is better on this board.Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0, as of Oct 28, 2026. Source | $0.012Cost per task | 6 |
| OpenRouter · τ²-bench airline: time per taskListed as: Claude Haiku 5.5via Anthropic, its own providerLower is better on this board.Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0, as of Oct 28, 2026. Source | 90 sTime per task | 5 |
OpenRouter's list price per 1M tokens in each of our snapshots.
Unchanged across our 41 snapshots.
| Snapshot | Input $/1M | Output $/1M |
|---|---|---|
| Oct 28, 09:22 UTC | $0.1 | $0.5 |
| Oct 27, 16:03 UTC | $0.1 | $0.5 |
| Oct 27, 04:03 UTC | $0.1 | $0.5 |
| Oct 26, 16:04 UTC | $0.1 | $0.5 |
| Oct 26, 04:03 UTC | $0.1 | $0.5 |
| Oct 25, 16:04 UTC | $0.1 | $0.5 |
| Oct 25, 06:03 UTC | $0.1 | $0.5 |
| Oct 24, 15:04 UTC | $0.1 | $0.5 |
| Oct 24, 03:03 UTC | $0.1 | $0.5 |
| Oct 23, 15:03 UTC | $0.1 | $0.5 |
| Oct 23, 03:04 UTC | $0.1 | $0.5 |
| Oct 22, 15:03 UTC | $0.1 | $0.5 |
| Oct 22, 05:04 UTC | $0.1 | $0.5 |
| Oct 21, 15:03 UTC | $0.1 | $0.5 |
| Oct 21, 03:02 UTC | $0.1 | $0.5 |
| Oct 20, 15:04 UTC | $0.1 | $0.5 |
| Oct 20, 03:03 UTC | $0.1 | $0.5 |
| Oct 19, 15:04 UTC | $0.1 | $0.5 |
| Oct 19, 10:50 UTC | $0.1 | $0.5 |
| Oct 18, 15:03 UTC | $0.1 | $0.5 |
| Oct 18, 03:05 UTC | $0.1 | $0.5 |
| Oct 17, 15:03 UTC | $0.1 | $0.5 |
| Oct 17, 03:03 UTC | $0.1 | $0.5 |
| Oct 16, 15:02 UTC | $0.1 | $0.5 |
| Oct 16, 03:03 UTC | $0.1 | $0.5 |
| Oct 15, 17:02 UTC | $0.1 | $0.5 |
| Oct 15, 03:02 UTC | $0.1 | $0.5 |
| Oct 14, 15:03 UTC | $0.1 | $0.5 |
| Oct 14, 03:02 UTC | $0.1 | $0.5 |
| Oct 13, 15:03 UTC | $0.1 | $0.5 |
| Oct 13, 05:03 UTC | $0.1 | $0.5 |
| Oct 12, 15:03 UTC | $0.1 | $0.5 |
| Oct 12, 03:04 UTC | $0.1 | $0.5 |
| Oct 11, 15:02 UTC | $0.1 | $0.5 |
| Oct 11, 03:03 UTC | $0.1 | $0.5 |
| Oct 10, 15:01 UTC | $0.1 | $0.5 |
| Oct 10, 03:04 UTC | $0.1 | $0.5 |
| Oct 9, 15:05 UTC | $0.1 | $0.5 |
| Oct 9, 03:04 UTC | $0.1 | $0.5 |
| Oct 8, 15:03 UTC | $0.1 | $0.5 |
| Oct 8, 03:05 UTC | $0.1 | $0.5 |
Source: OpenRouter, as of Oct 28, 09:22 UTC
This model in colour, against every other model on the Epoch Capabilities Index that OpenRouter prices.
1 model on the index has no OpenRouter price and is not plotted.
| Model | Epoch Capabilities Index | Sample task | Best value |
|---|---|---|---|
| Claude Opus 5.5Anthropic | 168.83 | $0.08 | Yes |
| GPT-6 AstraOpenAI | 168.81 | $0.20 | |
| GPT-6.1 SolOpenAI | 168.37 | $0.04 | Yes |
| Claude Fable 5.1Anthropic | 166.49 | $0.20 | |
| Gemini 4 ArgonGoogle | 166.45 | $0.04 | |
| Claude Sonnet 5.5Anthropic | 166.18 | $0.04 | |
| Claude Opus 5Anthropic | 163.94 | $0.10 | |
| GPT-6 SolOpenAI | 163.73 | $0.04 | |
| Claude Fable 5Anthropic | 163.42 | $0.20 | |
| GPT-5.5 ProOpenAI | 162.91 | $0.66 | |
| GPT-5.6 SolOpenAI | 162.79 | $0.08 | |
| Muse WonderMeta | 161.85 | $0.08 | |
| GPT-5.6 TerraOpenAI | 160.09 | $0.044 | |
| GPT-5.5OpenAI | 159.68 | $0.11 | |
| GPT-5.4 ProOpenAI | 159.57 | $0.66 | |
| Claude Haiku 5.5Anthropic | 159.32 | $0.002 | Yes |
| Claude Opus 4.8Anthropic | 159.01 | $0.10 | |
| Gemini 3.7 FlashGoogle | 157.97 | $0.015 | |
| Kimi K3MoonshotAI | 157.86 | $0.042 | |
| Gemini 3.8 FlashGoogle | 157.44 | $0.017 | |
| GPT-5.4OpenAI | 157.31 | $0.055 | |
| Muse Spark 1.3Meta | 157.21 | $0.021 | |
| GPT-5.3-CodexOpenAI | 157.09 | $0.045 | |
| Claude Sonnet 5Anthropic | 156.88 | $0.04 | |
| Claude Opus 4.7Anthropic | 156.79 | $0.10 | |
| GPT-6 LunaOpenAI | 156.75 | $0.002 | |
| Grok 4.6xAI | 156.57 | $0.032 | |
| GPT-5.6 LunaOpenAI | 156.5 | $0.0044 | |
| GPT-5.2 ProOpenAI | 156.33 | $0.55 |
Our calculation: Epoch AI's index against the sample-task price (undefined tokens in, undefined out, at each model's OpenRouter list price; the formula is under the price chart).
Source: Epoch AI (CC BY), as of Oct 28, 2026Source
Source: OpenRouter, as of Oct 28, 09:19 UTC
Epoch AI (CC BY) logged new runs: Claude Haiku 5.5 at 0.896 on GPQA diamond and 0.238 on SimpleQA Verified. GPT-6 Astra leads GPQA diamond at 0.958. OpenRouter lists Haiku 5.5 at $0.1 in / $0.5 out per 1M.
Why it matters: Independent numbers for Anthropic's cheapest new model: close to the leaders on science questions, far behind on factual recall.
From Drop #015 · Sat, Oct 10 · Morning