Ranked by tau3-Banking - 17 models
Best AI models for tool use
In board data through August 4, 2026, Qwen3.8-Max ranks first on the official tau3-Banking (tau-knowledge) leaderboard with 55.2%, 6.5 points ahead of Claude Opus 5 at 48.7%. Its list price is $2 input / $6 output per 1M tokens. A fixed rule, the lowest output price per 1M tokens among models with a registry price that score within 5 points of the top score, picks the leader itself. The highest-placed open-weight model is Kimi K3, ranked 11th with 37.1%.
Board data through , checked . How this is ranked
tau3-Banking (tau-knowledge) ranking
| Rank | Model | Provider | tau3-Banking | Source | Input / 1M | Output / 1M | Context | Weights | tau2-benchthru Mar 2, 2026 |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Qwen3.8-MaxBoard: Qwen 3.8 Max (Qwen) | Alibaba Cloud (Qwen) | 55.2%Official leaderboard - tau3-Banking leaderboard (default tab on taubench.com), banking_knowledge pass^1; standard submission by Sierra; reasoning effort: xhigh; retrieval config: alltools; user simulator gpt-5.2; rank #1 of standard submissions | taubench.com | $2 | $6 | 1M | Closed | - |
| 2 | Claude Opus 5Board: Claude Opus 5 (Anthropic) | Anthropic | 48.7%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: max; retrieval config: alltools; user simulator gpt-5.2; rank #2 | taubench.com | $5 | $25 | 1M | Closed | - |
| 3 | Grok 4.5Board: Grok 4.5 (xAI) | xAI | 47.9%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: high; retrieval config: alltools; user simulator gpt-5.2; rank #3 | taubench.com | $2 | $6 | 500K | Closed | - |
| 4 | GPT-5.6 SolBoard: GPT-5.6-sol (OpenAI) | OpenAI | 46.9%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: xhigh; retrieval config: alltools; user simulator gpt-5.2; rank #4 | taubench.com | $4 | $20 | 1.05M | Closed | - |
| 5 | GPT-5.5Board: GPT-5.5 (OpenAI) | OpenAI | 44.6%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: xhigh; retrieval config: alltools; user simulator gpt-5.2; rank #5 | taubench.com | $5 | $30 | 1.05M | Closed | - |
| 6 | Muse Spark 1.1 (Meta) | - | 40.5%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: xhigh; retrieval config: alltools; user simulator gpt-5.2; rank #6 | taubench.com | - | - | - | - | - |
| 7 | Claude Opus 4.7Board: Claude Opus 4.7 (Anthropic) | Anthropic | 40.2%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: max; retrieval config: alltools; user simulator gpt-5.2; rank #7 | taubench.com | $5 | $25 | 1M | Closed | - |
| 8 | Claude Opus 4.8Board: Claude Opus 4.8 (Anthropic) | Anthropic | 39.7%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: max; retrieval config: alltools; user simulator gpt-5.2; rank #8 (tied 39.7 with Claude Fable 5) | taubench.com | $5 | $25 | 1M | Closed | - |
| 8 | Claude Fable 5Board: Claude Fable 5 (Anthropic) | Anthropic | 39.7%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: max; retrieval config: alltools; user simulator gpt-5.2; rank #9 (tied 39.7 with Claude Opus 4.8) | taubench.com | $10 | $50 | 1M | Closed | - |
| 10 | GPT-5.4Board: GPT-5.4 (OpenAI) | OpenAI | 39.4%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: xhigh; retrieval config: alltools; user simulator gpt-5.2; rank #10 | taubench.com | $2.50 | $15 | 1.05M | Closed | - |
| 11 | Kimi K3Board: Kimi K3 (Moonshot AI) | Moonshot AI | 37.1%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: max; retrieval config: alltools; user simulator gpt-5.2; rank #11 (tied 37.1 with GLM-5.2) | taubench.com | $3 | $15 | 1.05M | Open | - |
| 11 | GLM-5.2Board: GLM-5.2 (Z.ai) | Z.ai | 37.1%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: xhigh; retrieval config: alltools; user simulator gpt-5.2; rank #12 (tied 37.1 with Kimi K3) | taubench.com | $1.40 | $4.40 | 1M | Open | - |
| 13 | GPT-5.2Board: GPT-5.2 (OpenAI) | OpenAI | 32.2%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: high; retrieval config: alltools; user simulator gpt-5.2; rank #13 | taubench.com | $1.75 | $14 | 400K | Closed | 84.8% |
| 14 | Gemini 3 Flash PreviewBoard: Gemini 3 Flash (Google) | 27.3%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: high; retrieval config: terminal; user simulator gpt-5.2; rank #14 (tied 27.3 with Claude Opus 4.6) | taubench.com | $0.50 | $3 | 1.05M | Closed | 83.5% | |
| 14 | Claude Opus 4.6Board: Claude Opus 4.6 (Anthropic) | Anthropic | 27.3%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: max; retrieval config: alltools; user simulator gpt-5.2; rank #15 (tied 27.3 with Gemini 3 Flash) | taubench.com | $5 | $25 | 1M | Closed | - |
| 16 | Gemini 3.1 Pro PreviewBoard: Gemini 3.1 Pro Preview (Google) | 26.0%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: high; retrieval config: alltools; user simulator gpt-5.2; rank #16 | taubench.com | $2 | $12 | 1.05M | Closed | - | |
| 17 | Claude Sonnet 4.5Board: Claude Sonnet 4.5 (Anthropic) | Anthropic | 25.3%Official leaderboard - tau3-Banking leaderboard, banking_knowledge pass^1; standard submission by Sierra; reasoning effort: enabled; retrieval config: terminal; user simulator gpt-5.2; rank #17 | taubench.com | $3 | $15 | 200K | Closed | 76.4% |
Prices in USD per 1M tokens at the standard rate, from each provider's pricing page. A dash means the board name does not match a model in our registry or the model is not on that board. Each related-board score links to its source.
Related
API capabilities
Providers of the top 3
Curated picks
FAQ
What is the best AI model for tool use?
Qwen3.8-Max ranks first by tau3-Banking (tau-knowledge), board data through August 4, 2026, with 55.2% (official leaderboard). One benchmark is not a verdict: settings differ between rows, so test on your own workload before you switch.
What is the cheapest AI model for tool use near the top of tau3-Banking?
Taking every ranked model within 5 points of the top score that has a price in our registry, Qwen3.8-Max has the lowest output price: $2 input / $6 output per 1M tokens, with 55.2% by tau3-Banking (tau-knowledge), board data through August 4, 2026.
What is the best open-weight model for tool use?
Kimi K3 is the highest-placed open-weight model by tau3-Banking (tau-knowledge), board data through August 4, 2026, ranked 11th with 37.1%.
How is this tool use ranking built?
Rows follow the tau3-Banking (tau-knowledge) board's own order, checked on October 17, 2026; its newest entry is dated August 4, 2026, and a use case is only published while that date is within 120 days of the check. Each model appears once. Prices, context windows and weights come from each provider's first-party pricing page in our model registry; a model the registry does not list shows a dash.
Newsletter
Get new benchmark results as they land
One email per week. New model releases, the leaderboard moves that matter and what they mean for what you build.