{"slug": "mercury-2-5-llm-hits-770-tokens-per-second", "title": "Mercury 2.5 LLM hits 770 tokens per second", "summary": "Inception's Mercury 2.5 reasoning model scores 12 on the Artificial Analysis Intelligence Index and runs at 770 output tokens per second, according to Artificial Analysis. The model, released September 8, 2026, supports text input and output with a 260k-token context window and is priced at $0.25 per 1M input tokens and $0.75 per 1M output tokens, averaging $0.06 per task on the Intelligence Index. Mercury 2.5 generated 35M tokens during the Intelligence Index evaluation, below the 84M median.", "body_md": "# Mercury 2.5 Intelligence, Performance & Price Analysis\n\n### Model summary\n\nMercury 2.5 is above average in intelligence and well priced when comparing to other models of similar price. It's also notably fast and fairly concise. The model supports text input, outputs text, and has a 260k tokens context window.\n\nMercury 2.5 scores 12 on the Artificial Analysis Intelligence Index, placing it above average among comparable models (median: 12). When evaluating the Intelligence Index, it generated 35M tokens, which is fairly concise in comparison to the median of 84M.\n\nPricing for Mercury 2.5 is $0.25 per 1M input tokens (moderately priced, median: $0.25) and $0.75 per 1M output tokens (moderately priced, median: $0.90). On average, it costs $0.06 per task to evaluate Mercury 2.5 on the Intelligence Index.\n\nAt 770 tokens per second, Mercury 2.5 is notably fast (110).\n\n| Reasoning | Yes This page shows the reasoning version of this model. A non-reasoning variant may also exist. | \n|---|---|\n| Input modality | Supports: text | \n| Output modality | Supports: text | \n| Context window | 260k ~390 A4 pages of size 12 Arial font | \n\nMetrics are compared against models of the same class:\n\n- Non-reasoning models → compared only with other non-reasoning models\n- Reasoning models → compared across both reasoning and non-reasoning\n- Open weights models → compared only with other open weights models of the same size class:\n- Tiny: ≤4B parameters\n- Small: 4B–40B parameters\n- Medium: 40B–150B parameters\n- Large: >150B parameters\n- Proprietary models → compared across proprietary and open weights models of the same price range, using a blended 3:1 input/output price ratio:\n- <$0.15 per 1M tokens\n- $0.15–$1 per 1M tokens\n- >$1 per 1M tokens\n\nHighlights\n\n## IntelligenceUpdated\n\n### [Artificial Analysis Intelligence Index](https://artificialanalysis.ai/evaluations/artificial-analysis-intelligence-index)\n\n### Artificial Analysis Intelligence Index by Open Weights / Proprietary\n\nMeasures the performance of models on specific capabilities and industries\n\n### [Artificial Analysis Finance & Accounting Index](https://artificialanalysis.ai/models/capabilities/finance-and-accounting)\n\n### Intelligence Evaluations\n\n[AA-Briefcase v1.1](https://artificialanalysis.ai/evaluations/aa-briefcase)Updated\n\nAgentic knowledge work, (Elo-500)/2000\n\n[GDPval-AA v2.1](https://artificialanalysis.ai/evaluations/gdpval-aa)Updated\n\nAgentic real-world work tasks, (Elo-500)/2000\n\n[AutomationBench-AA](https://artificialanalysis.ai/evaluations/automationbench-aa)Updated\n\nAgentic SaaS workflows\n\nAgentic coding & terminal use\n\nCoding\n\nReasoning & knowledge\n\n[GDP.pdf](https://artificialanalysis.ai/evaluations/gdp-pdf)New\n\nProfessional document reasoning, All-pass\n\n[CritPt](https://artificialanalysis.ai/evaluations/critpt)Under review\n\nPhysics reasoning\n\nKnowledge\n\n1 - hallucination rate\n\nLong context reasoning\n\nLegal agentic work, criterion pass rate\n\nAgentic business operations\n\nQuantitative analysis on spreadsheets & documents\n\nAgentic tool use\n\nKubernetes incident root-cause analysis\n\nVisual reasoning\n\n[MLCR-AA](https://artificialanalysis.ai/evaluations/mlcr-aa)New\n\nMedical long context reasoning\n\n### AA-Briefcase v1.1Updated\n\n### AA-Briefcase Elo\n\n### AA-Omniscience\n\n### AA-Omniscience Index\n\n## Intelligence Index Comparisons\n\n### Intelligence Index vs. Cost per Intelligence Index Task\n\n## Token Use\n\n### Output Tokens per Intelligence Index Task\n\n## Cost\n\n### Cost per Intelligence Index Task\n\n### Cost to Run Artificial Analysis Intelligence Index\n\n### Pricing: Cache Hit, Input, and Output\n\n## Context Window\n\n### Context Window\n\n## Speed\n\nMeasured by Output Speed (tokens per second)\n\n### Output Speed\n\n### Time per Intelligence Index Task\n\n## Latency\n\nMeasured by Time (seconds) to First Token\n\n### Latency: Time To First Answer Token\n\n## End-to-End Response Time\n\nSeconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed\n\n### End-to-End Response Time\n\n## Frequently Asked Questions\n\nCommon questions about Mercury 2.5\n\nMercury 2.5 was released on September 8, 2026.\n\nMercury 2.5 was created by Inception.\n\nMercury 2.5 scores 12 on the Artificial Analysis Intelligence Index, placing it above average among other reasoning models in a similar price tier (median: 12).\n\nMercury 2.5 generates output at 770.4 tokens per second (based on Inception's API), which is well above average compared to other reasoning models in a similar price tier (median: 110.3 t/s).\n\nMercury 2.5 has a time to first token (TTFT) of 2.91s (based on Inception's API), which is somewhat higher than average compared to other reasoning models in a similar price tier (median: 2.22s).\n\nMercury 2.5 costs $0.25 per 1M input tokens (better than average, median: $0.25) and $0.75 per 1M output tokens (better than average, median: $0.90), based on Inception's API.\n\nMercury 2.5 costs $0.25 per 1M input tokens and $0.75 per 1M output tokens (based on Inception's API). For a blended rate (7:2:1 cache hit/input/output ratio), this is $0.14 per 1M tokens. Pricing may vary by provider. [Compare provider pricing](https://artificialanalysis.ai/models/mercury-2-5/providers)\n\nWhen evaluated on the Intelligence Index, Mercury 2.5 generated 35M output tokens, which is better than average compared to other reasoning models in a similar price tier (median: 84M).\n\nYes, Mercury 2.5 is a reasoning model. It uses extended thinking or chain-of-thought reasoning to work through complex problems before providing an answer.\n\nMercury 2.5 supports text input.\n\nMercury 2.5 supports text output.\n\nNo, Mercury 2.5 does not support image input. It can only process text.\n\nNo, Mercury 2.5 is not multimodal. It only supports text input.\n\nMercury 2.5 has a context window of 260k tokens. This determines how much text and conversation history the model can process in a single request.\n\nNo, Mercury 2.5 is proprietary. The model weights are not publicly available.\n\nMercury 2.5 is a proprietary model and Inception has not disclosed the model size or parameter count.\n\nMercury 2.5 achieves a score of 12 on the Artificial Analysis Intelligence Index. This composite benchmark evaluates models across reasoning, knowledge, mathematics, and coding.\n\nYes, Mercury 2.5 is available via API through 1 provider. [Compare API providers](https://artificialanalysis.ai/models/mercury-2-5/providers)\n\nMercury 2.5 is available through 1 API provider. [Compare providers](https://artificialanalysis.ai/models/mercury-2-5/providers)", "url": "https://wpnews.pro/news/mercury-2-5-llm-hits-770-tokens-per-second", "canonical_source": "https://artificialanalysis.ai/models/mercury-2-5", "published_at": "2026-09-23 22:16:19+00:00", "updated_at": "2026-09-23 22:30:01.862421+00:00", "lang": "en", "topics": ["large-language-models", "artificial-intelligence", "ai-research"], "entities": ["Mercury 2.5", "Inception", "Artificial Analysis", "Artificial Analysis Intelligence Index"], "alternates": {"html": "https://wpnews.pro/news/mercury-2-5-llm-hits-770-tokens-per-second", "markdown": "https://wpnews.pro/news/mercury-2-5-llm-hits-770-tokens-per-second.md", "text": "https://wpnews.pro/news/mercury-2-5-llm-hits-770-tokens-per-second.txt", "jsonld": "https://wpnews.pro/news/mercury-2-5-llm-hits-770-tokens-per-second.jsonld"}}