BFCL Multi-Turn Long Context

Long Context

Source ↗ Higher is better

Rank	Model	Provider	Score
1	xLAM 2 70b fc r	Salesforce	76.0%
2	Anthropic: Claude Opus 4.5	Anthropic	70.5%
3	xLAM 2 8b fc r	Salesforce	67.0%
4	GLM 4.6 (FC thinking)	Z.ai	66.5%
5	Gemini 3 Pro Preview	Google	64.0%
6	o3-2025-04-16	OpenAI	63.0%
7	grok-4-1-fast-reasoning	SpaceXAI	62.5%
8	Anthropic: Claude Sonnet 4.5	Anthropic	59.0%
9	BitAgent Bounty 8B	Bittensor	57.0%
10	xLAM 2 32b fc r	Salesforce	56.5%
11	Anthropic: Claude Haiku 4.5	Anthropic	56.0%
12	qwen3-235b-a22b-instruct-2507	Alibaba	55.5%
13	Moonshotai Kimi K2 Instruct	Moonshot AI	55.0%
14	Grok 4.1 Fast (Non-reasoning)	SpaceXAI	52.0%
15	Grok 4	SpaceXAI	50.5%
16	Command A Reasoning	Cohere	48.5%
18	Nanbeige4 3B Thinking 2511	Nanbeige	47.0%
19	GPT 5.2 2025 12 11	OpenAI	46.5%
20	ToolACE 2 8B	Huawei	46.0%
21	o4-mini-2025-04-16	OpenAI	45.5%
22	xLAM 2 3b fc r	Salesforce	45.5%
23	gpt-4.1-2025-04-14	OpenAI	43.0%
24	Qwen: Qwen3 32B	Alibaba	43.0%
25	Arch Agent 32B	Katanemo	41.5%
26	Nanbeige3.5 Pro Thinking	Nanbeige	41.0%
27	Qwen: Qwen3 30B A3B Instruct 2507	Alibaba	41.0%
28	gpt-4.1-mini-2025-04-14	OpenAI	40.0%
29	GPT 5 nano 2025 08 07	OpenAI	38.0%
30	Google: Gemini 2.5 FlashRetiring	Google	35.5%
31	DeepSeek: DeepSeek V3.2 ExpHistorical	DeepSeek	35.0%
17	↳DeepSeek V3.2 Exp (Prompt + Thinking)	DeepSeek	48.5%
32	Qwen: Qwen3 8BHistorical	Alibaba	34.5%
33	GPT 5 mini 2025 08 07	OpenAI	33.0%
34	Qwen: Qwen3 14B	Alibaba	32.5%
35	gpt-4.1-nano-2025-04-14	OpenAI	30.0%
36	Arch Agent 3B	Katanemo	29.0%
37	Meta: Llama 3.3 70B Instruct	Meta	26.5%
38	xLAM 2 1b fc r	Salesforce	25.5%
39	Qwen3 4B Instruct 2507	Alibaba	25.5%
40	Cohere: Command A	Cohere	25.0%
41	Arch Agent 1.5B	Katanemo	22.0%
42	Hammer2.1 7b	MadeAgents	21.5%
43	mistral-large-2411	Mistral AI	19.0%
44	Llama 4 Maverick 17B 128E Instruct FP8	Meta	18.0%
45	Google: Gemini 2.5 Flash LiteRetiring	Google	17.5%
46	Hammer2.1 1.5b	MadeAgents	16.5%
47	Hammer2.1 3b	MadeAgents	15.5%
48	Google: Gemma 3 27B	Google	14.0%
49	mistral-medium-2505	Mistral AI	13.5%
50	Meta: Llama 3.1 8B Instruct	Meta	13.0%

Access this data

JSON GET /api/v1

[
  {
    "rank": 1,
    "model": "xLAM 2 70b fc r",
    "provider": "Salesforce",
    "score": "76.0"
  },
  {
    "rank": 2,
    "model": "Anthropic: Claude Opus 4.5",
    "provider": "Anthropic",
    "score": "70.5"
  },
  {
    "rank": 3,
    "model": "xLAM 2 8b fc r",
    "provider": "Salesforce",
    "score": "67.0"
  }
]

Curl

$ curl -s https://rightrank.ai/api/v1/rankings/bfcl-multi-turn-long-context.json | jq '.[:3]'

CLI brew install rightrank

$ rightrank top bfcl-multi-turn-long-context --limit 5
$ rightrank compare claude-opus-4.6 gpt-5.4-pro --category bfcl-multi-turn-long-context
$ rightrank value bfcl-multi-turn-long-context  # best performance per dollar

Claude Code Skill Skill

# In Claude Code, use the RightRank skill:
/rightrank best model for bfcl-multi-turn-long-context
/rightrank compare top 5 by value
/rightrank cheapest model scoring above 80 on bfcl-multi-turn-long-context