Commit History
28 days ago
-
Add GPT 6.1 Sol
676755c -
Add Claude Opus 5.5, Claude Sonnet 5.5, GPT 6 Sol, Astra and Luna; update prices
1ec3cbe -
Open source the eval harness
7fd8268
-
Add cost/point over time
3fcd299
-
Unify build script and add more models
406eae5
-
Add Qwen 3.8 27B (XHigh, Q4_K_XL)
c0ff6a9
-
Add Qwen 3.8 27B (Medium, Q4_K_XL)
890124c
-
Update generatedAt
2f833d2 -
Add Gemini 3.7 Flash (Medium)
e0eb74c -
Fix incorrect release year
d75a060 -
Add Deepseek V4 Pro 0813 (High)
2e10e8d
-
Add Qwen 3.8 2.4T-A95B
a4782f1
-
Add Grok 4.6 (Medium)
496bac5 -
Add Nemotron 3.5 Lightning
af4e4f8
-
Add topic breakdown and details for Muse Glimmer 30B (Q4_K_XL)
3d4ff5c -
Add Muse Glimmer 30B (Q4_K_XL)
9c04c4a -
Update topic methodology to match new UI
44d4222 -
Major refactor and results explorer
25cfbfb
-
Add Deepseek V4 Flash 0731
988702a
-
Rerun GPT 5.6 Luna/Terra at new pricing
89a790d
-
Add topic breakdown for Qwen2.5-Coder-32B
51f08f9 -
Add Qwen 2.5 Coder 32B, Qwen 3 4B Instruct 2507, Deepseek R1 0528, and Devstral Small 2507
bbd8973
-
Add Qwen 3 Coder 30B-A3B, Claude Sonnet 4, Gemini 2.5 Pro, Gemini 3.1 Pro, GPT 5
3bca641
-
Display a lot more information
778b7a2 -
Breakdown QA scores by topic
afbc698 -
Add fail/noCode to leetcode results and the exact release date of the model
d138419 -
Start cost animation at constant offset
372ac63 -
Improve chart zoom/pan responsiveness
3f0ee19 -
Add chart zoom and adjust the copy
a079f1c -
Rerun benchmarks with more robust retries for fewer no_answer failures
dfbe9cd -
Improve chart readability
a085767 -
Add provider logos to make charts easier to read
a666225
-
Adjust price according to OpenRouter effective pricing
c534a41 -
Add avg noAnswer to overall table view
8e00fc1 -
Better breakdowns and table layout
47ff52c -
Add Kimi K3
43d0a9b -
More precise data entry with less rounding
d291a2e -
Add Claude Opus 5
ca980fd
-
Add total eval costs
c5a42ff -
Add Grok 4.5 (Medium)
cb4281e
-
Add GPT 5.6
58de8fb
-
Add Qwen 3.5 9B
4217ba2
-
Mention the number of knowledge questions
a6655b5 -
Use the same rounding as everyone else
b7dce9d -
Improve tie behavior
30419b5 -
Add Claude Fable 5
cf883f4 -
Add llama 3.1 8b
98a3e2e
-
Initial commit
b7dfece