The AI Model Scout
Two leaderboards: every model ranked on intelligence per dollar, and on how well it searches the web, against what it actually costs to run a task.
GPT-6.1 Sol (Xhigh)
Best price-to-quality
- Intelligence
- 51.0
- Web search
- 41.0
Cheapest model that still keeps near-frontier intelligence
Most intelligence per dollarGLM-5.3-Flash
Best for web search
- Intelligence
- 41.8
- Web search
- 53.0
Smartest web researcher for the least money
Best at web searchBest value right now
GPT-6.1 Sol (Xhigh)
OpenAICheapest model that still keeps near-frontier intelligence
- Intelligence
- 51.0
- Cost/task
- $0.39
- Input
- $2.00/M tokens
- Output
- $10.00/M tokens
- Speed
- 51 tok/s
- Web search
- 41.0
- Value score
- 129.9
Why this one
GPT-6.1 Sol (Xhigh) scores 51.0 on the Intelligence Index, 89% of the leader Claude Opus 5.5 (Max, Default Fallback), while costing $0.39 per task. That is 15x cheaper than the leader. It sits on the Pareto frontier, so no model is both smarter and cheaper.
Best for web search right now
GLM-5.3-Flash
Z AISmartest web researcher for the least money
- Web search
- 53.0
- Intelligence
- 41.8
- Cost/task
- $0.28
- Context
- 1M
- Most accurate (least hallucination)
- 72%
- Speed
- 65 tok/s
Why this one
GLM-5.3-Flash scores 53.0 on the deep-research benchmark, the highest on the board, while keeping 41.8 general intelligence and costing only $0.28 per task. For search-heavy work it beats far pricier models at a fraction of the cost.
What changed
Since the previous scan
New models (3)
- Llama 4 Maverick(Meta)
- Gemma 4 E4B (Reasoning)(Google)
- Llama 4 Scout(Meta)
Cost per task moved (2)
- Ling 3.1 Flash: $1.00 → $1.00 per task
- Qwen3.8 2.4T A95B: $2.16 → $2.16 per task
Most intelligence per dollar
Ranked by Intelligence Index against what one full task actually costs.
| Model | Provider | Intelligence | Cost/task | Input | Output | Speed | Value score |
|---|---|---|---|---|---|---|---|
| Claude Opus 5.5 (Max, Default Fallback)Best in class | Anthropic | 57.6 | $5.98 | $4.00 | $20.00 | 95 | 9.6 |
| Claude Sonnet 5.5 (Max, Default Fallback)Best in class | Anthropic | 56.0 | $5.46 | $2.00 | $10.00 | 133 | 10.3 |
| Claude Opus 5.5 (Xhigh, Default Fallback)Best in class | Anthropic | 56.0 | $3.46 | $4.00 | $20.00 | 76 | 16.2 |
| Claude Opus 5.5 (High, Default Fallback)Best in class | Anthropic | 53.6 | $1.82 | $4.00 | $20.00 | 77 | 29.4 |
| Claude Fable 5.1 (Max, Default Fallback)Claude Opus 5.5 (Max, Default Fallback) is smarter and cheaper | Anthropic | 53.4 | $7.63 | $10.00 | $50.00 | 71 | 7.0 |
| Claude Fable 5.1 (Xhigh, Default Fallback)Claude Sonnet 5.5 (Max, Default Fallback) is smarter and cheaper | Anthropic | 53.2 | $5.98 | $10.00 | $50.00 | 63 | 8.9 |
| GPT-6 Astra (Max)Claude Opus 5.5 (High, Default Fallback) is smarter and cheaper | OpenAI | 52.7 | $3.26 | $10.00 | $50.00 | 44 | 16.2 |
| Gemini 4 Argon (High)Claude Opus 5.5 (High, Default Fallback) is smarter and cheaper | 52.6 | $1.99 | $2.00 | $10.00 | not published | 26.4 | |
| Muse Wonder (Max)Claude Opus 5.5 (Xhigh, Default Fallback) is smarter and cheaper | Meta | 52.4 | $4.52 | $4.00 | $20.00 | 154 | 11.6 |
| GPT-6 Astra (Xhigh)Claude Opus 5.5 (High, Default Fallback) is smarter and cheaper | OpenAI | 52.4 | $2.31 | $10.00 | $50.00 | 44 | 22.7 |
| Claude Sonnet 5.5 (Xhigh, Default Fallback)Claude Opus 5.5 (High, Default Fallback) is smarter and cheaper | Anthropic | 51.9 | $2.01 | $2.00 | $10.00 | 100 | 25.8 |
| GPT-6.1 Sol (Max)Best in class | OpenAI | 51.8 | $0.72 | $2.00 | $10.00 | 49 | 71.6 |
| Claude Opus 5.5 (Medium, Default Fallback)GPT-6.1 Sol (Max) is smarter and cheaper | Anthropic | 51.2 | $1.34 | $4.00 | $20.00 | 86 | 38.4 |
| Claude Fable 5.1 (High, Default Fallback)Claude Opus 5.5 (Xhigh, Default Fallback) is smarter and cheaper | Anthropic | 51.1 | $3.91 | $10.00 | $50.00 | 59 | 13.1 |
| GPT-6.1 Sol (Xhigh)Best price-to-quality | OpenAI | 51.0 | $0.39 | $2.00 | $10.00 | 51 | 129.9 |
| GPT-6 Astra (High)GPT-6.1 Sol (Max) is smarter and cheaper | OpenAI | 50.9 | $1.73 | $10.00 | $50.00 | 47 | 29.5 |
| GPT-6.1 Sol (High)Best in class | OpenAI | 50.2 | $0.32 | $2.00 | $10.00 | 51 | 157.4 |
| GPT-6 Astra (Medium)GPT-6.1 Sol (Max) is smarter and cheaper | OpenAI | 49.6 | $1.54 | $10.00 | $50.00 | 49 | 32.2 |
| Claude Fable 5.1 (Medium, Default Fallback)Claude Opus 5.5 (High, Default Fallback) is smarter and cheaper | Anthropic | 48.9 | $2.98 | $10.00 | $50.00 | 57 | 16.4 |
| Muse Spark 1.3 (Max)GPT-6.1 Sol (Max) is smarter and cheaper | Meta | 48.1 | $1.60 | $1.25 | $4.25 | 128 | 30.0 |
Showing 20 of 116 models
Best at web search
Ranked by the deep-research benchmark: how well each model searches and synthesizes the web.
| Model | Provider | Web search | Intelligence | Cost/task | Input | Output | Context |
|---|---|---|---|---|---|---|---|
| Gemini 4 Argon (High) | 63.5 | 52.6 | $1.99 | $2.00 | $10.00 | 1M | |
| MiniMax-M3 | MiniMax | 60.0 | 29.2 | $0.51 | $0.30 | $1.20 | 1M |
| Muse Wonder (Max) | Meta | 57.5 | 52.4 | $4.52 | $4.00 | $20.00 | 1M |
| Grok 4.7 (Xhigh) | SpaceXAI | 53.0 | 46.5 | $3.74 | $2.00 | $6.00 | 500K |
| GLM-5.3-FlashBest researcher | Z AI | 53.0 | 41.8 | $0.28 | $0.15 | $0.50 | 1M |
| Qwen3.8 Max (0902) | Alibaba | 51.5 | 45.4 | $5.41 | $2.00 | $6.00 | 984K |
| GLM-5.3 (Max) | Z AI | 51.5 | 44.8 | $2.01 | $1.40 | $4.40 | 1M |
| Muse Spark 1.3 (Max) | Meta | 51.0 | 48.1 | $1.60 | $1.25 | $4.25 | 1M |
| Muse Spark 1.3 (Xhigh) | Meta | 50.5 | 45.1 | $1.37 | $1.25 | $4.25 | 1M |
| Grok 4.7 (High) | SpaceXAI | 49.5 | 46.3 | $2.73 | $2.00 | $6.00 | 500K |
| Claude Sonnet 5.5 (Max, Default Fallback) | Anthropic | 49.0 | 56.0 | $5.46 | $2.00 | $10.00 | 1M |
| Ling 3.1 Flash | InclusionAI | 48.5 | 41.1 | $1.00 | $0.30 | $0.90 | 1M |
| MiMo-V2.6-Pro | Xiaomi | 47.5 | 46.3 | $0.13 | $0.43 | $0.87 | 1M |
| Claude Haiku 5.5 (Max) | Anthropic | 47.5 | 43.4 | $0.21 | $0.10 | $0.50 | 1M |
| Grok 4.7 (Low) | SpaceXAI | 47.0 | 42.2 | $1.25 | $2.00 | $6.00 | 500K |
| Qwen3.8 2.4T A95B | Alibaba | 47.0 | 40.0 | $2.16 | $2.00 | $6.00 | 262K |
| Qwen3.7 Plus | Alibaba | 47.0 | 25.2 | $0.22 | $0.40 | $1.60 | 1M |
| Nemotron 3 Ultra 550B A55B (Reasoning) | NVIDIA | 46.5 | 22.9 | $0.60 | $0.60 | $2.50 | 262K |
| GPT-6 Astra (High) | OpenAI | 46.0 | 50.9 | $1.73 | $10.00 | $50.00 | 1M |
| Step 5 Preview | StepFun | 46.0 | 43.7 | $1.03 | $1.00 | $2.70 | 1M |
Showing 20 of 111 models
Token prices are USD per million tokens. Cost per task is the USD cost of one full Intelligence Index run. Speed is output tokens per second. Web search is the deep-research benchmark score, and context is the maximum window in tokens.
At a glance
- Highest intelligence:
- Claude Opus 5.5 (Max, Default Fallback) (57.6)
- Best at web search:
- GLM-5.3-Flash
- Most accurate (least hallucination):
- Command A+ (86%)
- Fastest output:
- Muse Wonder (Max)
- Lowest cost per task:
- GPT-6 Luna (Low)
- Average cost per task:
- $1.04
- On the Pareto frontier:
- 17 of 116 models
How this is ranked
Value score is Intelligence Index divided by cost per task. The best value pick is the cheapest model still scoring at least 46.1, which is 80% of the leader Claude Opus 5.5 (Max, Default Fallback). Figures come straight from the source's published datasets, not from an AI summary.