Hy3
Output 94 tokens/sFirst token 1.9 sContext 256K tokens
Top 3 cheapest providers (Adjusted $/task)
The same list price can give a different adjusted $/task (caching, token efficiency) — click a price for its inputs.
Within the active global provider, residency and confidentiality filters.
| # | Provider | Price channel | Raw input $/1M | Raw output $/1M | Adjusted $/task |
|---|---|---|---|---|---|
| 1 | Tencentvia OpenRouter — Price via OpenRouter: OpenRouter's price for Tencent's endpoint (https://openrouter.ai/tencent/hy3/providers), collected 2026-11-03. | via OpenRouter — Price via OpenRouter: OpenRouter's price for Tencent's endpoint (https://openrouter.ai/tencent/hy3/providers), collected 2026-11-03. | $0.132 | $0.528 | |
| 2 | Novitavia OpenRouter — Price via OpenRouter: OpenRouter's price for Novita's endpoint (https://openrouter.ai/tencent/hy3/providers), collected 2026-11-03. | via OpenRouter — Price via OpenRouter: OpenRouter's price for Novita's endpoint (https://openrouter.ai/tencent/hy3/providers), collected 2026-11-03. | $0.140 | $0.580 | |
| 3 | Phalavia OpenRouter — Price via OpenRouter: OpenRouter's price for Phala's endpoint (https://openrouter.ai/tencent/hy3/providers), collected 2026-11-03. | via OpenRouter — Price via OpenRouter: OpenRouter's price for Phala's endpoint (https://openrouter.ai/tencent/hy3/providers), collected 2026-11-03. | $0.150 | $0.640 |
Composite
2 of 7 inputs6 radar axes: DesignArena's two boards share one36.7
Thin data · 2/7Based on only 2 of 7 Composite inputs — treat this rank as uncertaindominance-adjusted from 38.7: a better-measured model that is at least as good on each of these inputs ranks above it
AA Coding v4.x 58.8Coding Agent v1.4 —AA Intelligence (est.) 12.7AA Agentic 25.6Epoch ECI —Software ECI —DesignArena —/—
Radar: percentile among all models measured on each input; a gap means not measured.
Benchmark sheet
16 of 319 registered benchmark versions · bars show the percentile among all models measured on each benchmark.
Agentic
AA-Briefcase50950 Elo
AA-Briefcase published 2026-09-10 → · Published board — Tests multi-week professional knowledge-work projects with linked tasks and large source collections.
GDPval-AA v2591,136 Elo
GDPval-AA v2 v2 → · Published board — Tests professional knowledge-work deliverables across occupations using AA's Stirrup harness.
Terminal-Bench v2.1 (AA)6364.4%
Terminal-Bench v2.1 (AA) v2.1 → · Published board — Tests terminal-based work on the 89-task verified refresh using Terminus 2.
Terminal-Bench v4.0 (AA)400.5%
Terminal-Bench v4.0 (AA) v4.0 → · Published board — Tests terminal-based work on the 66-task release using mini-SWE-agent v2.4.6.
Long-Horizon Terminal-Bench v1.0200.29
Long-Horizon Terminal-Bench v1.0 → · Terminus-2 — An agent works in a terminal for up to 90 minutes per task on 46 long jobs — building systems, migrating frameworks, playing games move by move — and earns partial credit for real progress.
AA Agentic Index6725.6
AA Agentic Index published 2026-09-18 → · Published board — Artificial Analysis publishes one Agentic Index per model, measured on its primary configuration and relayed by OpenRouter's Benchmarks API; the value is attached at family scope on the deterministic representative. Artificial Analysis publishes this board without a version number; we keep the date each result was retained.
Coding
SciCode (AA subproblems) v1.0.14948.6%
SciCode (AA subproblems) v1.0.1 → · Published board — Tests scientific Python programming with scientist-annotated background information.
AA Coding Index v4.x · frozen 2026-10-276858.8
AA Coding Index v4.x · frozen 2026-10-27 → · Published board — Artificial Analysis withdrew the Coding Index with Intelligence Index v5.0 (2026-10-28). These are the last v4.x values, frozen with their own date (2026-10-27).
Knowledge
Humanity's Last Exam (AA text-only)8333.5%
Humanity's Last Exam (AA text-only) published 2026-09-10 → · Published board — Tests expert-level knowledge on AA's text-only Humanity's Last Exam subset.
AA-Omniscience Index63-18.5
AA-Omniscience Index published 2026-09-10 → · Published board — Tests factual reliability while rewarding correct answers and penalizing hallucinations.
Long-context
GDP.pdf (AA)5011.2%
GDP.pdf (AA) published 2026-09-10 → · Published board — Tests professional reasoning over long PDFs with AA document preparation and grading.
AA-LCR v1.18479.0%
AA-LCR v1.1 v1.1 → · Published board — Tests reasoning across multiple long documents with corrected answer keys and grading.
Context Arena — MRCR v2 (8-needle)426.9%
Context Arena — MRCR v2 (8-needle) 8-needle → · Published board — The model must find and reproduce the right one of eight near-identical earlier replies hidden in a long conversation, averaged over context lengths from 8k to 128k tokens.
Roleplay
Japanese-RP-Bench v2 (tegnike fork)8094.6
Japanese-RP-Bench v2 (tegnike fork) v2 → · Published board — Fork of Japanese-RP-Bench adding v2 metrics for role fidelity, conversation quality, persona stability, robustness, and recovery in Japanese roleplay.
Science
CritPt (AA)774.9%
CritPt (AA) published 2026-09-10 → · Published board — Tests research-level physics reasoning with Python, symbolic and numerical answers.
GPQA Diamond (AA)8989.7%
GPQA Diamond (AA) published 2026-09-10 → · Published board — Tests graduate-level biology, physics and chemistry knowledge on the Diamond subset.
Tool-use
AutomationBench-AA v1.0.64218.4%
AutomationBench-AA v1.0.6 → · Published board — Tests multi-app SaaS workflows through REST tools on a held-out AutomationBench split.
τ³-Banking (AA) v1.0.15822.9%
τ³-Banking (AA) v1.0.1 → · Published board — Tests banking support agents that retrieve policies and change account state through tools.
Unusual results
1 threshold-crossing signal flagged · 15 eligible benchmark families
How flags are calculated
Heuristic screen, not statistical significance: benchmark families are correlated and source uncertainty is unknown. Peer evidence needs ≥ 20 independently measured matched configurations from ≥ 10 distinct model families. A flag needs a directed population z-score of magnitude ≥ 1.5 and a gap of ≥ 1.5 from the leave-one-benchmark-family-out mean z in the same direction, over ≥ 5 other benchmark families.
- unusually weak Context Arena — MRCR v2 (8-needle)
Why
Observed: 26.94049 percentPeer mean 73.38237 · peer sd 21.36553 · n 47 · families 31Directed z -2.174 · baseline z 0.356 · gap -2.53 · profile n 1426.94049 percentderived from measuredobserved 2026-09-21contextarena.ai ↗Evidence
Axis: Context Arena — MRCR v2 (8-needle) · 8-needle · Published boardExact value:26.94049033136195percentObserved: 2026-09-21T17:13:26.665692+00:00 · publication date: not recordedObservation id:public:9a01d34c2c07108bef0d42d9
Missing coverage · 303 benchmark versions
No result does not mean a zero, or that the model was never tested. Collection failures and disputed versions retain their distinct status.
Token offers by platform · 3 offers (Adjusted $/task)
Click any underlined price to see how it is estimated and where each input comes from. How we calculate adjusted cost.
3 offers within the active global filters, 2 more outside them; “price n/a” means the provider serves the model but publishes no direct token price we could collect — no other channel's price is shown in its place.
OpenRouter (3)
Prices via OpenRouter — what OpenRouter charges for each provider's endpoint, not the provider's own direct price. Source: openrouter.ai/tencent/hy3/providers, collected 2026-11-03.
| Tencentconditional | globaltencent/fp8 | $0.132 raw in $/1M | $0.528 raw out $/1M | |
| Novita | globalnovita | $0.140 raw in $/1M | $0.580 raw out $/1M | |
| Phala | globalphalaTEE | $0.150 raw in $/1M | $0.640 raw out $/1M |